An experimental data processing method and apparatus

Abstract

Embodiments of the present application disclose an experimental data processing method and device; after obtaining an experimental index data set, according to the experimental grouping and data type of the experimental index data in the experimental index data set, an experimental data set to be processed for the experimental index is constructed, then, based on a preset individual difference parameter, a difference regression coefficient of each experimental index is identified in the experimental data set to be processed, panel experimental data is screened out in the experimental data set to be processed, and based on the panel experimental data, the variance of the difference regression coefficient is calculated to obtain difference detection data of each experimental index, then, when receiving a difference detection request for a target experimental index, difference detection is performed on the target experimental index according to the difference detection data, and the detection result is sent to a terminal; the scheme can improve the processing efficiency of experimental data processing.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of communication, in particular to an experimental data processing method and device. BACKGROUND

[0002] Due to the randomness of the shunting algorithm, the inconsistency between the shunted crowd and the strategy reached crowd, the non-standard experiment operation, the shunting algorithm deviation and other reasons, the experimental indicators may have A / A difference, thereby affecting the evaluation of the experimental effect. In order to correct the difference of the experimental indicators, it is often necessary to process the experimental data, and the existing experimental data processing method usually adopts the difference correction method of double difference algorithm (DID).

[0003] In the research and practice process of the prior art, the inventors of the present application found that different experimental objects are active in different experimental stages, and in the collected mixed data, there may be a situation that part of the data of the experimental objects before or after the experiment is "missing". For this kind of mixed data, DID algorithm cannot be directly used for difference correction, thereby affecting the efficiency of difference detection, and thus the processing efficiency of experimental data processing is low. SUMMARY

[0004] The experimental data processing method and device provided by the embodiments of the present application can improve the processing efficiency of experimental data processing.

[0005] An experimental data processing method, comprising:

[0006] Obtaining an experimental index data set, the experimental index data set comprising experimental index data of at least one experimental index of at least one experimental object under a current experiment;

[0007] According to the experimental grouping and data type of the experimental index data, a to-be-processed experimental data set for the experimental index is constructed;

[0008] Based on a preset individual difference parameter, a difference regression coefficient of each experimental index is identified in the to-be-processed experimental data set;

[0009] Panel data is screened out in the to-be-processed experimental data set, and based on the panel data, the variance of the difference regression coefficient is calculated to obtain difference detection information of each experimental index;

[0010] When receiving a difference detection request for a target experimental index, difference detection is performed on the target experimental index according to the difference detection information, and the detection result is sent to a terminal.

[0011] Optionally, the experimental data processing method provided by the embodiments of the present application can further comprise:

[0012] sending a difference detection request for a target experiment index to a server;

[0013] receiving a detection result of the target experiment index returned by the server;

[0014] displaying an experiment data processing page, the experiment data processing page including a difference significance display control corresponding to a type of the detection result;

[0015] in response to a display operation on the difference significance display control, displaying the detection result of the target experiment index.

[0016] Correspondingly, an experiment data processing apparatus is provided in an embodiment of the present application, comprising:

[0017] an acquisition unit configured to acquire an experiment index data set, the experiment index data set including experiment index data of at least one experiment index of at least one experiment object under a current experiment;

[0018] a construction unit configured to construct a to-be-processed experiment data set for the experiment index according to experiment grouping and data types of the experiment index data;

[0019] an identification unit configured to identify a difference regression coefficient of each experiment index in the to-be-processed experiment data set based on a preset individual difference parameter;

[0020] a calculation unit configured to filter out panel experiment data in the to-be-processed experiment data set, and calculate a variance of the difference regression coefficient based on the panel experiment data, to obtain difference detection information of each experiment index;

[0021] a detection unit configured to, when receiving a difference detection request for a target experiment index, perform difference detection on the target experiment index according to the difference detection information, and send a detection result to a terminal.

[0022] Optionally, an experiment data processing apparatus is also provided in an embodiment of the present application, comprising:

[0023] a sending unit configured to send a difference detection request for a target experiment index to a server;

[0024] a receiving unit configured to receive a detection result of the target experiment index returned by the server;

[0025] a first display unit configured to display an experiment data processing page, the experiment data processing page including a difference significance display control corresponding to a type of the detection result;

[0026] The second display unit is configured to display a detection result of the target experimental index in response to a display operation on the difference significance display control.

[0027] Optionally, in some embodiments, the identifying unit can be specifically configured to aggregate an intermediate statistical quantity of the experimental index from the to-be-processed experimental data set according to the experimental grouping and the data type; identify a data region with missing experimental index data in the to-be-processed experimental data set; filter data in the data region to obtain a target experimental data set; and identify a difference regression coefficient of each experimental index in the target experimental data set according to the preset individual difference parameter and the intermediate statistical quantity.

[0028] Optionally, in some embodiments, the identifying unit can be specifically configured to obtain a region quantity of a data region in the target experimental data set; extract an experimental index feature and an experimental parameter feature from the target experimental data set; and fuse the experimental index feature and the experimental parameter feature based on the intermediate statistical quantity, the preset individual difference parameter and the region quantity to obtain the difference regression coefficient of each experimental index.

[0029] Optionally, in some embodiments, the identifying unit can be specifically configured to perform feature conversion on the experimental parameter feature to obtain a converted experimental parameter feature; fuse the converted experimental parameter feature and the experimental parameter feature based on the region quantity and the preset individual difference parameter to obtain a fused experimental parameter feature; fuse the converted experimental parameter feature and the experimental index feature based on the preset individual difference parameter to obtain a fused experimental index feature; and determine the difference regression coefficient of each experimental index based on the intermediate statistical quantity, the fused experimental parameter feature and the fused experimental index feature.

[0030] Optionally, in some embodiments, the identifying unit can be specifically configured to fuse the fused experimental parameter feature and the fused experimental index feature based on the intermediate statistical quantity to obtain difference regression information; and extract the difference regression coefficient of each experimental index from the difference regression information.

[0031] Optionally, in some embodiments, the identifying unit can be specifically configured to classify experimental index data in the to-be-processed experimental data set according to the data type; aggregate an initial intermediate statistical quantity of the experimental index from each type of experimental index data; and fuse the initial intermediate statistical quantities to obtain the intermediate statistical quantity of the experimental index.

[0032] Optionally, in some embodiments, the computing unit can be specifically configured to extract residual features of the corresponding experimental indicators from the panel data according to the difference regression coefficient; split pre-experiment data and post-experiment data from the to-be-processed experimental data set; calculate the variance of the difference regression coefficient of the corresponding experimental indicators based on the intermediate statistical quantity, the residual features, the pre-experiment data and the post-experiment data; and take the difference regression coefficient and the variance of the difference regression coefficient as the difference detection information of the experimental indicators.

[0033] Optionally, in some embodiments, the computing unit can be specifically configured to filter out pre-experiment indicator data and post-experiment indicator data corresponding to each experimental indicator from the panel data; determine residual information of the corresponding experimental indicators based on the difference regression coefficient, the pre-experiment indicator data and the post-experiment indicator data; and extract residual features of the experimental indicators from the residual information.

[0034] Optionally, in some embodiments, the computing unit can be specifically configured to obtain preset pre-experiment parameter information and preset post-experiment parameter information; fuse the difference regression coefficient with the preset pre-experiment parameter information to obtain fused pre-experiment data, and fuse the difference regression coefficient with the preset post-experiment parameter information to obtain fused post-experiment data; calculate a residual between the pre-experiment indicator data and the fused pre-experiment data to obtain pre-experiment residual, and calculate a residual between the post-experiment indicator data and the fused post-experiment data to obtain post-experiment residual; and take the pre-experiment residual and the post-experiment residual as residual information of the corresponding experimental indicators.

[0035] Optionally, in some embodiments, the computing unit can be specifically configured to extract pre-experiment features from the pre-experiment data and extract post-experiment features from the post-experiment data; fuse the residual features with the pre-experiment features to obtain fused pre-experiment features, and fuse the residual features with the post-experiment features to obtain fused post-experiment features; and fuse the fused pre-experiment features and the fused post-experiment features based on the intermediate statistical quantity to obtain the variance of the difference regression coefficient of the corresponding experimental indicators.

[0036] Optionally, in some embodiments, the detection unit can be specifically configured to filter out target difference detection information corresponding to the target experimental indicator from the difference detection information; calculate difference data of the target experimental indicator according to the target difference detection information; and determine a detection result of the target experimental indicator based on the difference data.

[0037] Optionally, in some embodiments, the detection unit can be specifically configured to extract a target difference regression coefficient and a target variance of the target difference regression coefficient from the target difference detection information; identify a post-experiment difference regression coefficient from the target difference regression coefficient, and fuse the post-experiment difference regression coefficient and an experiment cumulative day number to obtain a modified post-experiment difference information; calculate a difference significance probability of the target experiment index based on the target difference regression coefficient and the target variance, and take the modified post-experiment difference information and the difference significance probability as difference data.

[0038] Optionally, in some embodiments, the detection unit can be specifically configured to fuse the target difference regression coefficient and a target variance to obtain a target detection statistic of the target experiment index; acquire basic probability information of a detection statistic not exceeding the target detection statistic, and calculate target probability information of the detection statistic exceeding the target detection statistic based on the basic probability information; determine difference significance probability information of the target experiment index based on the target probability information; and extract a pre-experiment difference significance probability and a modified post-experiment difference significance probability from the difference significance probability information as the difference significance probability of the target experiment index.

[0039] Optionally, in some embodiments, the detection unit can be specifically configured to extract a pre-experiment difference significance probability from the difference data; when the pre-experiment difference significance probability does not exceed a preset probability threshold, take the difference data as a detection result of the target experiment index; and when the pre-experiment difference significance probability exceeds the preset probability threshold, perform difference detection on post-experiment experiment index data corresponding to the target experiment index to obtain a detection result.

[0040] Optionally, in some embodiments, the acquisition unit can be specifically configured to acquire at least one current experiment index data of at least one experiment object at a current experiment day; update corresponding experiment index cumulative data based on the current experiment index data; calculate a daily average cumulative value of each updated experiment index cumulative data of the experiment object to obtain an experiment index data set.

[0041] Optionally, in some embodiments, the first display unit can be specifically configured to, when a type of the detection result is existence of a pre-experiment difference, extract a modified post-experiment difference significance probability from the detection result, and display an experiment data processing page based on the modified post-experiment difference significance probability; and when the type of the detection result is non-existence of a pre-experiment difference, acquire a difference significance display control corresponding to a post-experiment difference, and display an experiment data processing page based on the difference significance display control corresponding to the post-experiment difference.

[0042] Optionally, in some embodiments, the first display unit can be specifically used for determining that the correction type of the target experimental index is a correction significant type when the corrected experimental difference significance probability does not exceed the preset significance probability threshold, and acquiring a difference significance display control corresponding to the correction significant type to display the experimental data processing page; and determining that the correction type of the target experimental index is a correction insignificant type when the corrected experimental difference significance probability exceeds the preset significance probability threshold, and acquiring a difference significance display control corresponding to the correction insignificant type to display the experimental data processing page.

[0043] In addition, an electronic device is also provided in the embodiments of the present application, which comprises a processor and a memory, the memory stores an application program, and the processor is used to run the application program in the memory to implement the experimental data processing method provided in the embodiments of the present application.

[0044] In addition, a computer readable storage medium is also provided in the embodiments of the present application, which stores a plurality of instructions, the instructions are suitable for being loaded by a processor to execute the steps in any one of the experimental data processing methods provided in the embodiments of the present application.

[0045] After obtaining the experimental index data set, the embodiments of the present application construct a to-be-processed experimental data set for the experimental index according to the experimental grouping and data type of the experimental index data in the experimental index data set, then identify the difference regression coefficient of each experimental index in the to-be-processed experimental data set based on a preset individual difference parameter, filter out panel experimental data in the to-be-processed experimental data set, and calculate the variance of the difference regression coefficient based on the panel experimental data to obtain the difference detection information of each experimental index, then when receiving the difference detection request for the target experimental index, perform difference detection on the target experimental index according to the difference detection information, and send the detection result to the terminal. Since the preset difference parameter is introduced in the scheme, the correlation of the sample data before and after the experiment is fully considered, not only the panel data and the mixed cross-sectional data can be processed, but also the mixed data composed of the panel data and the mixed cross-sectional data can be processed, the different experimental indexes of the same experiment can be detected and processed at the same time, thereby reducing the consumption of computing resources, and therefore the processing efficiency of experimental data processing can be improved. BRIEF DESCRIPTION OF DRAWINGS

[0046] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed in the embodiment description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0047] Figure 1 is a schematic diagram of an experimental data processing system provided by an embodiment of the present application;

[0048] Figure 2 is a schematic diagram of a scene of an experimental data processing method provided by an embodiment of the present application;

[0049] Figure 3 is a schematic diagram of a flow of an experimental data processing method provided by an embodiment of the present application;

[0050] Figure 4 is a schematic diagram of panel data and mixed cross sections provided by an embodiment of the present application;

[0051] Figure 5 is a schematic diagram of an intermediate statistic of a pre-polymerization experimental index provided by an embodiment of the present application;

[0052] Figure 6 is a schematic diagram of a flow of difference detection on an experimental index provided by an embodiment of the present application;

[0053] Figure 7 is a schematic diagram of a comparison result corresponding to simulation 1 provided by an embodiment of the present application;

[0054] Figure 8 is a schematic diagram of a comparison result corresponding to simulation 2 provided by an embodiment of the present application;

[0055] Figure 9 is a schematic diagram of a comparison result corresponding to simulation 3 provided by an embodiment of the present application;

[0056] Figure 10 is a schematic diagram of a comparison result corresponding to simulation 4 provided by an embodiment of the present application;

[0057] Figure 11 is another schematic diagram of a flow of an experimental data processing method provided by an embodiment of the present application;

[0058] Figure 12 is a page diagram of an experimental data processing page provided by an embodiment of the present application;

[0059] Figure 13 is another schematic diagram of a flow of an experimental data processing method provided by an embodiment of the present application;

[0060] Figure 14 is a structural schematic diagram of a first experimental data processing device provided by an embodiment of the present application;

[0061] Figure 15 is a structural schematic diagram of a second experimental data processing device provided by an embodiment of the present application;

[0062] Figure 16is a structural schematic diagram of an electronic device provided by an embodiment of the present application. DETAILED DESCRIPTION

[0063] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by a person skilled in the art without creative work fall within the protection scope of the present application.

[0064] The present application provides an experimental data processing method and device. Wherein, the experimental data processing device can be integrated in an electronic device, which can be a server, a terminal or the like. Specifically, the present application provides an experimental data processing device suitable for a first electronic device (which can be referred to as a first experimental data processing device for distinction) and an experimental data processing device suitable for a second electronic device (which can be referred to as a second experimental data processing device for distinction).

[0065] Wherein, the first electronic device can be a network side device such as a server, which can be a stand-alone physical server, a server cluster or a distributed system composed of multiple physical servers, or a cloud server providing cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, network acceleration services (CDN), and big data and artificial intelligence platforms, etc. The second electronic device can be a terminal, which can be a smart phone, a tablet computer, a notebook computer, a desktop computer, a smart speaker, a smart watch, etc., but is not limited thereto. The terminal and the server can be directly or indirectly connected through wired or wireless communication, which is not limited in the present application.

[0066] The present application takes the first electronic device as a server and the second electronic device as a terminal as an example to introduce the experimental data processing method.

[0067] For example, referring to Figure 1 The present application provides an experimental data processing system including a server 10 and a terminal 20, which are connected through a network, such as a wired or wireless network connection, etc. Wherein, the experimental data processing device can be integrated in the terminal, such as in the form of a client.

[0068] The server 10 can be configured to, after obtaining the experimental index data set, construct a to-be-processed experimental data set for the experimental index according to the experimental grouping and data type of the experimental index data in the experimental index data set, then identify a difference regression coefficient of each experimental index in the to-be-processed experimental data set based on a preset individual difference parameter, filter out panel experimental data in the to-be-processed experimental data set, calculate the variance of the difference regression coefficient based on the panel experimental data, obtain difference detection information of each experimental index, and then, when receiving a difference detection request for a target experimental index, perform difference detection on the target experimental index according to the difference detection information, and send the detection result to the terminal, thereby improving the processing efficiency of experimental data processing, such as Figure 2 as shown.

[0069] The terminal 20 can send a difference detection request for a target experimental index to the server and receive a detection result of the target experimental index returned by the server. Specifically, the terminal 20 can perform the following operations:

[0070] The terminal 20 can send a difference detection request for a target experimental index to the server and receive a detection result of the target experimental index returned by the server. Specifically, the terminal 20 can perform the following operations:

[0071] Wherein, in response to the condition or state that the operation performed depends on, when the dependent condition or state is met, the one or more operations performed can be real-time or have a set delay; in the absence of specific instructions, there is no restriction on the execution order of the multiple operations performed.

[0072] It can be understood that in the specific embodiments of the present application, the experimental index data of the experimental object and other related data are involved. When the following embodiments of the present application are applied to specific products or technologies, the user's permission or consent is required, and the collection, use and processing of related data need to comply with relevant laws, regulations and standards of relevant countries and regions.

[0073] The following are described in detail. It should be noted that the order of the following embodiments is not limited as the preferred order of the embodiments.

[0074] The embodiment will be described from the perspective of a first experimental data processing apparatus, which can be integrated in an electronic device, which can be a server, which can be a standalone physical server, a server cluster or a distributed system formed by multiple physical servers, or a cloud server providing cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, content delivery network (CDN), and basic cloud computing services such as big data and artificial intelligence platforms.

[0075] An experimental data processing method, comprising:

[0076] Obtaining an experimental index data set, the experimental index data set comprising experimental index data of at least one experimental index of at least one experimental object under a current experiment, constructing a to-be-processed experimental data set for the experimental index according to the experimental grouping and data type of the experimental index data, identifying a difference regression coefficient of each experimental index in the to-be-processed experimental data set based on a preset individual difference parameter, filtering panel experimental data in the to-be-processed experimental data set, and calculating the variance of the difference regression coefficient based on the panel experimental data to obtain difference detection information of each experimental index, when receiving a difference detection request for a target experimental index, performing difference detection on the target experimental index according to the difference detection information, and sending the detection result to a terminal.

[0077] As shown in Figure 3 , the specific process of the experimental data processing method is as follows:

[0078] 101、Obtaining an experimental index data set.

[0079] The experimental index data set comprises experimental index data of at least one experimental index of at least one experimental object under a current experiment.

[0080] The experimental index data set can be obtained in various ways, which can be as follows:

[0081] For example, at least one current experimental index data of at least one experimental object under a current experimental day can be obtained, the corresponding experimental index cumulative data is updated based on the current experimental index data, and the ratio of each updated experimental index cumulative data to the experimental cumulative day is calculated to obtain the experimental index data set.

[0082] The current experimental index data can be understood as the experimental index data collected on the next day of the current experimental day. For example, if the current experimental day is the second day, the current experimental index data can be understood as the experimental index data collected on the second day. After obtaining the current experimental index data, the corresponding experimental index cumulative data can be updated based on the current experimental index data. The updating method can be various, such as obtaining the experimental index cumulative data of each experimental index, querying the target experimental index cumulative data corresponding to each current experimental index data in the experimental index cumulative data according to the experimental index corresponding to the current experimental index data, adding the current experimental index data to the target experimental index cumulative data when the target experimental index cumulative data exists, obtaining the updated experimental index cumulative data, and when the target experimental index cumulative data does not exist, taking the current experimental index data as the updated experimental index cumulative data. Alternatively, the experimental index data of multiple days can be directly added to obtain the updated experimental index cumulative data.

[0083] After updating the experimental index cumulative data, the daily average cumulative value of each updated experimental index cumulative data of the experimental object can be calculated. The daily average cumulative value can be understood as the ratio of the experimental index data of the experimental object collected in the observation period to the cumulative experimental days. Therefore, the method for calculating the daily average cumulative value can be various, such as obtaining the experimental cumulative days corresponding to each updated experimental index cumulative data of the experimental object, calculating the ratio between each updated experimental index cumulative data and the corresponding experimental cumulative days, thereby obtaining the experimental index data of each experimental index of the experimental object, and combining the experimental index data of each experimental object to obtain the experimental index data set.

[0084] 102. Constructing a to-be-processed experimental data set for the experimental index according to the experimental grouping and data type of the experimental index data.

[0085] The experimental grouping is used to indicate the grouping information of the experimental index data of the experimental object in the experimental scene. The experimental grouping can include various types, such as experimental group and control group.

[0086] The data type can be understood as the collection time of the experimental index data of the experimental object in the experimental scene. The collection time can include various types, such as before the experiment and after the experiment.

[0087] The method for constructing the to-be-processed experimental data set for the experimental index according to the experimental grouping and data type of the experimental index data can be various, which can be as follows:

[0088] For example, the attribute information of each experimental index data in the experimental index data set can be acquired, the experimental group and the data type of the experimental index data are identified in the attribute information, the experimental grouping parameter of the experimental index data is determined based on the experimental group, the data grouping parameter of the experimental index data is determined based on the data type, the experimental type parameter and the data type parameter are fused to obtain the experimental index fusion parameter, and the experimental index fusion parameter, the experimental grouping parameter, the data type parameter and the experimental index data are combined to obtain the to-be-processed experimental data set.

[0089] For example, when the experimental group is an experimental group, the corresponding experimental grouping parameter A can be 1 or other preset parameter value, and when the experimental group is a control group, the corresponding experimental grouping parameter A can be 0 or other preset parameter value.

[0090] For example, when the data type is pre-experimental data, the corresponding data type parameter T can be 0 or other preset parameter value, and when the data type is post-experimental data, the corresponding data type parameter T can be 1.

[0091] For example, when the experimental grouping parameter A is 0 or 1 and the data type parameter T is 0 or 1, the to-be-processed experimental data set for the experimental index can be as shown in Table 1:

[0092] Table 1

[0093]

[0094]

[0095] As can be seen from Table 1, the experimental index data processed in the present scheme can include a mixture of panel data and mixed cross-sectional data. The so-called panel data can be understood as experimental index data of active experimental objects before and after the experiment. The characteristic of the panel data is that the observation data of the same experimental object at different time points is not independent, but the data between different experimental objects can be considered to be independent. The so-called mixed cross-sectional data can be understood as experimental index data of active experimental objects only before or after the experiment. The characteristic of the mixed cross-sectional data is that it does not contain repeated experimental objects, and the data can be considered to be independent of each other. The panel data and the mixed cross-sectional data can be as shown in Table 2. Figure 4

[0096] 103、Based on the preset individual difference parameter, the difference regression coefficient of each experimental index in the to-be-processed experimental data set is identified.

[0097] ​wherein the preset individual difference parameter (a i ) is used to measure the correlation between the data of the same experimental subject at different observation periods.

[0098] wherein the difference regression coefficient can also be understood as a DID regression coefficient of the experimental index, and is used to measure the difference between the experimental group and the control group before the experiment, the difference after the experiment after correction, etc., and can be used to calculate the test statistics of each regression coefficient.

[0099] wherein the way of identifying the difference regression coefficient of each experimental index in the to-be-processed experimental data set can be various, and can be as follows:

[0100] For example, according to the data type and the experimental grouping, the intermediate statistics of the experimental index are aggregated in the to-be-processed experimental data set, the data region with missing experimental index data is identified in the to-be-processed experimental data set, the data in the data region is filtered out to obtain a target experimental data set, and according to the preset individual difference parameter and the intermediate statistics, the difference regression coefficient of each experimental index is identified in the target experimental data set.

[0101] wherein the intermediate statistics can be an intermediate statistical aggregation result required for calculating the regression coefficient of the experimental index and the variance of the regression coefficient, and the type of the intermediate statistics can be various, such as the intermediate aggregation result before the experiment (A / A period), the intermediate aggregation result after the experiment (A / B period), and the intermediate aggregation result of the panel data. The intermediate aggregation result before the experiment (A / A period) can include the A / A period sample quantity, the experimental pre-index data sum, and the experimental pre-index data square sum aggregated according to the experimental grouping. The intermediate aggregation result after the experiment (A / B period) can include the A / B period sample quantity, the experimental post-index data sum, and the experimental post-index data square sum aggregated according to the experimental grouping. The intermediate aggregation result of the panel data can include the panel data sample quantity, the experimental pre-index data sum, the experimental pre-index data square sum, the experimental post-index data sum, the experimental post-index data square sum, and the product sum of the experimental pre-index data and the experimental post-index data aggregated according to the experimental grouping. The way of aggregating the intermediate statistics of the experimental index in the to-be-processed experimental data set according to the experimental grouping and the data type can be various, such as classifying the experimental index data in the to-be-processed experimental data set according to the experimental grouping and the data type, respectively aggregating the initial intermediate statistics of the experimental index in each type of experimental index data set, and fusing the initial intermediate statistics to obtain the intermediate statistics of the experimental index.

[0102] There are several ways to classify experimental index data in the dataset based on data type. For example, based on data type, the mixed cross-sectional data and panel data extracted from the dataset can be divided into pre-experiment cross-sectional data and post-experiment cross-sectional data according to experimental grouping. Similarly, panel data can be divided into pre-experiment panel data and post-experiment panel data according to experimental grouping. The pre-experiment panel data and pre-experiment cross-sectional data can be used as pre-experiment sample data, and the post-experiment panel data and post-experiment cross-sectional data can be used as post-experiment sample data. Therefore, the experimental index data in the dataset can be classified into panel data, pre-experiment sample data, and post-experiment sample data.

[0103] After classifying the panel data, pre-experiment sample data, and post-experiment sample data, the initial intermediate statistics of the experimental indicators can be aggregated from these three types of experimental indicator data. There are various aggregation methods, such as aggregating the sample size of the panel data, the sample size of period A / A, and the sample size of period A / B, as well as aggregating the sum, sum of squares, and product of the pre-experiment and post-experiment indicator data, etc. The aggregated intermediate results are used as the initial intermediate statistics of the experimental indicators.

[0104] After aggregating the initial intermediate statistics of the experimental indicators, the initial intermediate statistics can be merged. There are several ways to merge them. For example, they can be directly summarized or spliced ​​together. Alternatively, they can obtain the weighting coefficient of each data type, and then weight the initial intermediate statistics based on the weighting coefficient. Finally, the weighted initial intermediate statistics are merged to obtain the intermediate statistics of each experimental indicator.

[0105] In this study, the design matrices for multiple experimental indicators within the same experiment can be reused. Therefore, in specific engineering deployments, data can be pre-aggregated in blocks to ensure faster and more flexible generation of intermediate statistics required for each experimental indicator in difference detection. Specifically, this can be achieved as follows: Figure 5 As shown, the experimental indicator data for each experimental subject is collected daily through the user indicator daily table. The user indicator cumulative table is updated based on the experimental indicator data of the day. Then, based on the experimental exposure table, sample data before the experiment (A / A period), sample data after the experiment (A / B period), and panel data are selected from the user indicator cumulative table. Intermediate statistics corresponding to the experimental indicators are pre-aggregated for different experimental data. These intermediate statistics may include the sample size, indicator sum, indicator square sum, and product sum of indicators before and after the experiment for different data regions aggregated according to the experimental group.

[0106] In the to-be-processed experimental data set, the data region where the experimental index data is missing can be identified in various ways. For example, in the to-be-processed experimental data set, the data set df shown in Table 1, the data region corresponding to the data row with null data can be regarded as the data region of the experimental index data.

[0107] After identifying the data region where the experimental index data is missing, the data in the data region can be filtered. The filtering method can be various, such as directly deleting other data in the data region, or directly deleting the data region in the to-be-processed experimental data set df to obtain the target experimental data set df filt .

[0108] After filtering the data in the data region to obtain the target experimental data set, the difference regression coefficient of each experimental index in the target experimental data set can be identified. The method of identifying the difference regression coefficient of each experimental index can be various, such as obtaining the number of data regions in the target experimental data set, extracting the experimental index feature and the experimental parameter feature in the target experimental data set, fusing the experimental index feature and the experimental parameter feature based on the intermediate statistical quantity, the preset individual difference parameter and the region number, to obtain the difference regression coefficient of each experimental index.

[0109] In which, the way of obtaining the region number of the data region in the target experimental data set can be various, such as identifying the number of rows N filt in the target experimental data set df filt , and regarding each row as a data region. Therefore, the number of rows N filt in df filt can be the region number of the data region.

[0110] In which, the experimental index feature can be understood as the characteristic information of the experimental index data, and the experimental parameter feature can be understood as the characteristic information of the experimental parameter. The experimental parameter can be the experimental grouping parameter (A), the data type parameter (T) and the experimental index fusion parameter (T*A). The way of extracting the experimental index feature and the experimental parameter feature in the target experimental data set can be various, such as extracting the experimental parameter in the target experimental data set to obtain the experimental parameter set, and extracting the experimental index data in the target experimental data set to obtain the target experimental index data set, extracting the features of the experimental parameter set to obtain the experimental parameter feature, and extracting the features of the target experimental index data set to obtain the experimental index feature.

[0111] In which, the way of extracting the features of the experimental parameter set can be various, such as constructing a first basic feature vector X filt= (1, T, A, T*A), taking the experimental parameters of each row in the experimental parameter set as elements in the first basic feature vector, so as to obtain the first feature vector corresponding to the experimental parameters, and taking the first feature vector as the experimental parameter feature.

[0112] Wherein, the way of feature extraction of the target experimental index data set can be various, such as, a second basic feature vector Y filt = (Y1,.....,Y K , taking the experimental index data of each experimental index in the target experimental index data set as elements in the second basic feature vector, so as to obtain the second feature vector corresponding to the experimental index, and taking the second feature vector as the experimental index feature.

[0113] After extracting the experimental index feature and the experimental parameter feature, the experimental index feature and the experimental parameter feature can be fused to obtain the difference regression coefficient of each experimental index. The way of fusing the experimental index feature and the experimental parameter feature can be various, such as, the experimental parameter feature can be transformed to obtain the transformed experimental parameter feature, based on the number of regions, the transformed experimental parameter feature and the experimental parameter feature are fused to obtain the fused experimental parameter feature, according to the individual difference parameter, the transformed experimental parameter feature and the experimental index feature are fused to obtain the fused experimental index feature, based on the intermediate statistical quantity, the fused experimental parameter feature and the fused experimental index feature, the difference regression coefficient of each experimental index is determined.

[0114] Wherein, the way of transforming the experimental parameter feature can be various, such as, taking the experimental parameter feature as the first feature vector X filt , just transposing the first feature vector X filt to obtain the transposed first feature vector , taking the transposed first feature vector as the transformed experimental parameter feature.

[0115] After transforming the experimental parameter feature, the transformed experimental parameter feature and the experimental parameter feature can be fused based on the number of regions and the preset individual difference parameter, and the way of fusion can be various, such as, based on N filt , the transposed first feature vector and the first feature vector X filt are multiplied, so as to obtain the fused experimental parameter feature, which can be specifically shown in formula (1):

[0116]

[0117] Wherein, is the fused experimental parameter feature, Nfilt is an experimental grouping parameter, T is a data type parameter, and (T·A) is an experimental index fusion parameter.

[0118] After the feature conversion of the experimental parameter feature, the converted experimental parameter feature and the experimental index feature can be fused according to a preset individual difference parameter to obtain a fused experimental index feature. The fusion manner can be various, for example, the converted first feature vector Y1 and the second basic feature vector Y2 can be multiplied based on the preset individual difference parameter to obtain the fused experimental index feature. Specifically, the fused experimental index feature can be as shown in formula (2): filt

[0119]

[0120] wherein, is the fused experimental index feature, Y1...Y K may be experimental index data of K experimental indexes, A is an experimental grouping parameter, T is a data type parameter, and (T·A) is an experimental index fusion parameter.

[0121] After the fused experimental parameter feature and the fused experimental index feature are obtained, the difference regression coefficient of each experimental index can be determined based on the intermediate statistical quantity, the fused experimental parameter feature and the fused experimental index feature. The manner of determining the difference regression coefficient can be various, for example, the fused experimental parameter feature and the fused experimental index feature are fused based on the intermediate statistical quantity to obtain difference regression information, and the difference regression coefficient of each experimental index is extracted from the difference regression information.

[0122] wherein the difference regression information can be understood as a parameter set composed of regression coefficients of each experimental index. The form of the set can be a matrix. Therefore, the manner of fusing the fused experimental parameter feature and the fused experimental index feature to obtain the difference regression information can be various, for example, the inverse matrix of the fused experimental parameter feature Y can be multiplied by the fused experimental index feature Y to obtain a difference regression matrix. Based on the element values corresponding to each element in the difference regression matrix calculated in the intermediate statistical quantity, the element values are replaced with the corresponding elements in the difference regression matrix to obtain the difference regression information. Specifically, the difference regression information can be as shown in formula (3):

[0123]

[0124] wherein, is the difference regression information, is the fused experimental parameter feature, ​​​​​To fuse the experimental index characteristics after the experiment.

[0125] Need to be noted that the data contained in the intermediate statistics can be understood as the sum, product, and square sum of the experimental index data generated when doing matrix calculation, thus, the element value corresponding to each element in the difference regression matrix can be directly calculated in the intermediate statistics, without complicated calculation, thereby improving the processing efficiency of experimental data processing.

[0126] After obtaining the difference regression information Then, the difference regression coefficient of each experimental index can be extracted from the difference regression information. The way of extracting the difference regression coefficient can be various, for example, each column in the difference regression information may represent the DID regression coefficient corresponding to each experimental index. The elements of each column can be taken as the difference regression coefficient of the experimental index, which can be shown as formula (3):

[0127]

[0128] wherein, is the difference regression coefficient of the Kth experimental index, correspond to different parameters in the difference detection model. The so-called difference detection model can be understood as a method of combining DID and random effect model to realize the A / A difference correction and variance reduction model. The model form is shown as formula (4):

[0129] Y ij = a0+ a i + a1T ij + a2A ij + a3A ij T ij + ∈ ij (4)

[0130] i = 1, …, n, j = 1, …, t i

[0131] T ij ∈ {0, 1}, A ij ∈ {0, 1},

[0132]

[0133] wherein, Y ij represents the observed experimental index (experimental index data), T ij represents whether it is data after the experiment (1 represents after the experiment, and 0 represents before the experiment), and A ijrepresents whether it is an experimental group (1 represents an experimental group, and 0 represents a control group), i represents the ith user, j represents the jth observation value, each experimental subject can have 1 or 2 observation experimental index data (t i is 1 or 2). By introducing the individual difference parameter a i , the correlation between the data of the same user at different observation periods is measured. The Kth experimental index a0, a1, a2, and a3 can be obtained respectively.

[0134] 104. Panel data is screened out from the to-be-processed experimental data set, and the variance of the difference regression coefficient is calculated based on the panel data to obtain difference detection data of each experimental index.

[0135] Among them, the panel data can be experimental index data of an experimental subject that is active before and after the experiment, and can also be understood as experimental index data of an experimental subject that is observed multiple times during the entire experimental period.

[0136] Among them, the difference detection data can be understood as detection data required for detecting A / A difference of an experimental index. Through the detection data, the pre-experiment difference significance probability, the corrected post-experiment difference significance probability, and the corrected post-experiment difference can be quickly calculated when detection is required. The difference detection data can mainly include the difference regression coefficient of each experimental index and the variance of the difference regression coefficient.

[0137] Among them, there can be multiple ways to screen out panel data from the to-be-processed experimental data set, which can be as follows:

[0138] For example, the data of experimental subjects with only one observation can be filtered out from the to-be-processed experimental data set to obtain panel data. For example, taking the data set df shown in Table 1 as an example, it can be found in Table 1 that experimental subjects c and d only have pre-experiment experimental index data, and experimental subjects e and f only have post-experiment experimental index data. Therefore, the data of the rows where experimental subjects c, d, e, and f are located can be filtered out, and panel data df both can be obtained. It can be found from Table 1 that the panel data can be experimental index data and experimental parameters of experimental subjects a and b.

[0139] After the panel data is screened out, the variance of the difference regression coefficient can be calculated based on the panel data. There can be multiple ways to calculate the variance of the difference regression coefficient, such as extracting the residual feature of the corresponding experimental index in the panel data according to the difference regression coefficient, splitting the pre-experiment data and post-experiment data in the to-be-processed experimental data set, calculating the variance of the difference regression coefficient of the corresponding experimental index data based on the intermediate statistical quantity, the residual feature, the pre-experiment data and the post-experiment data, and taking the difference regression coefficient and the variance of the difference regression coefficient as the difference detection information of the experimental index.

[0140] There can be multiple ways to extract the residual feature of the corresponding experimental index in the panel data according to the difference regression coefficient, such as screening out the pre-experiment experimental index data and post-experiment experimental index data corresponding to each experimental index in the panel data, determining the residual information of the corresponding experimental index based on the difference regression coefficient, the pre-experiment experimental index data and the post-experiment experimental index data, and extracting the residual feature of the experimental index in the residual information.

[0141] The pre-experiment experimental index data can be the experimental index data when the data type parameter T is 0, and the post-experiment experimental index data can be the experimental index data when the data type parameter T is 1. Therefore, there can be multiple ways to screen out the pre-experiment experimental index data and post-experiment index data corresponding to each experimental index in the panel data, such as screening out the experimental index data of each experimental object when the data type parameter T is 0 in the panel data to obtain the pre-experiment experimental index data, and screening out the experimental index data when the data type T is 1 in the panel data to obtain the post-experiment experimental index data. After screening out the pre-experiment experimental index data and the post-experiment index data, the data format of the panel data df both may be rearranged as use id ,A,Y 11 ,Y 12 ,......,Y K1 ,Y K2 。Y K1 represents the pre-experiment experimental index data, and Y K2 represents the post-experiment experimental index data.

[0142] After screening the pre-experiment index data and the post-experiment index data, the residual information of the corresponding experimental index can be determined based on the difference regression coefficient, the pre-experiment index data and the post-experiment index data. The residual information of the corresponding experimental index can be understood as the information indicating the residual covariance of the difference regression coefficient of the experimental index. There can be various ways to determine the residual information of the corresponding experimental index, for example, obtaining preset pre-experiment parameter information and preset post-experiment parameter information, fusing the difference regression coefficient with the preset pre-experiment parameter information to obtain fused pre-experiment data, fusing the difference regression coefficient with the preset post-experiment parameter information to obtain fused post-experiment data, calculating the residual of the pre-experiment index data and the fused pre-experiment data to obtain pre-experiment residual, and calculating the residual of the post-experiment index data and the fused post-experiment data to obtain post-experiment residual, taking the pre-experiment residual and the post-experiment residual as the residual information of the corresponding experimental index.

[0143] wherein the preset pre-experiment parameter information can be indicative of a pre-experiment feature vector before the experiment, and the pre-experiment feature vector can be X pre =(1, 0, A, 0), and the preset post-experiment parameter information corresponding thereto can be indicative of a post-experiment feature vector after the experiment, and the post-experiment feature vector can be X post =(1, 1, A, A).

[0144] wherein the way to calculate the residual of the pre-experiment index data and the fused pre-experiment data can be various, for example, directly calculating the data difference value of the pre-experiment index data and the fused pre-experiment data, taking the data difference value as the pre-experiment residual, which can be specifically shown in formula (5):

[0145]

[0146] wherein e K1 is the pre-experiment residual of the Kth experimental index, Y K1 is the pre-experiment index data of the Kth experimental index, X pre is the preset pre-experiment parameter information, is the difference regression coefficient of the Kth experimental index.

[0147] wherein the way to calculate the residual of the post-experiment index data and the fused post-experiment data can be various, for example, directly calculating the difference value of the post-experiment index data and the fused post-experiment data to obtain the post-experiment residual, which can be specifically shown in formula (6):

[0148]

[0149] wherein e K2 is the post-experiment residual of the Kth experimental index, Y K2X represents the post-experimental index data for the Kth experimental index. post To preset the parameter information after the experiment, is the difference regression coefficient for the Kth experimental indicator.

[0150] After calculating the pre-experiment residuals and post-experiment residuals, the pre-experiment residuals and post-experiment residuals can be used as the residual information of the corresponding experimental indicators.

[0151] After determining the residual information of the experimental index, the residual features of that experimental index can be extracted from the residual information. There are several ways to extract residual features; for example, the residual covariance matrix V of the residuals before and after the experiment can be directly calculated. K =Cov(e K1 ,e K2 ), and its inverse matrix W K =(Cov(e) K1 ,e K2 )) -1 The residual covariance matrix V K and its inverse matrix W K Residual characteristics as an experimental indicator.

[0152] There are several ways to divide the experimental data set into pre-experiment and post-experiment data. For example, taking the dataset df in Table 1 as the experimental data set to be processed, the data in the row with T=0 is taken as the pre-experiment data, and the data in the row with T=1 is taken as the post-experiment data. This can be understood as dividing the explanatory variables of df into two parts, pre-experiment and post-experiment, as shown in formula (7).

[0153]

[0154] Where X1 represents the data before the experiment, and X2 represents the data after the experiment.

[0155] After segmenting the pre-experiment and post-experiment data and extracting the residual features of the experimental indicators, the variance of the differential regression coefficients of the experimental indicators can be calculated. There are several ways to calculate the variance of the differential regression coefficients. For example, pre-experiment features can be extracted from the pre-experiment data and post-experiment features can be extracted from the post-experiment data. The residual features can be fused with the pre-experiment features to obtain the fused pre-experiment features, and the residual features can be fused with the post-experiment features to obtain the fused experimental features. Based on the intermediate statistics, the fused pre-experiment features and the fused post-experiment features can be fused to obtain the variance of the differential regression coefficients of the corresponding experimental indicators.

[0156] The pre-experiment features can be extracted from the pre-experiment data in various ways. For example, the basic pre-experiment features are extracted from the pre-experiment data, the basic pre-experiment features are converted to obtain converted basic pre-experiment features, the converted basic pre-experiment features are fused with the basic pre-experiment features to obtain first fused pre-experiment features, the basic post-experiment features are extracted from the post-experiment data, the converted basic pre-experiment features are fused with the basic post-experiment features to obtain second fused pre-experiment features, the first fused pre-experiment features are fused with the second fused pre-experiment features to obtain the pre-experiment features. The post-experiment features are extracted from the post-experiment data in the same way as the pre-experiment features. For example, the basic post-experiment features can be converted to obtain converted post-experiment features, the basic post-experiment features and the converted post-experiment features are fused to obtain first fused post-experiment features, the converted post-experiment features are fused with the basic pre-experiment features to obtain second fused post-experiment features, and the first fused post-experiment features and the second fused post-experiment features are fused to obtain the post-experiment features.

[0157] After the pre-experiment features and the post-experiment features are extracted, the residual features can be fused with the pre-experiment features and the post-experiment features, respectively. The fusion method can be various. For example, the pre-experiment residual features and the post-experiment residual features are cut from the residual features, and the cut pre-experiment residual features and the post-experiment residual features can be as shown in formula (8):

[0158]

[0159] wherein, W K is the inverse matrix of the residual covariance matrix corresponding to the Kth experimental index, W K11 and W K12 are pre-experiment residual features, W K21 and W K22 are post-experiment residual features.

[0160] The pre-experiment residual features are fused with the pre-experiment features to obtain fused pre-experiment features, and the post-experiment residual features are fused with the post-experiment features to obtain fused post-experiment features.

[0161] After obtaining the pre-fusion experiment features and the post-fusion experiment features, the pre-fusion experiment features and the post-fusion experiment features can be fused to obtain the variance of the difference regression coefficient of the corresponding experiment index. The fusion method can be various, for example, the pre-fusion experiment features and the post-fusion experiment features can be directly added to obtain the target variance feature. Based on the target element value of each element in the matrix corresponding to the target variance feature calculated in the intermediate statistical quantity, the target feature value of the target variance feature is determined based on the target element value, and the inverse matrix of the target feature value is calculated to obtain the variance of the difference regression coefficient of the experiment index. Specifically, it can be as shown in formula (9):

[0162]

[0163] wherein, is the variance of the difference regression coefficient of the Kth experiment index, W K11 and W K12 is the pre-experiment residual feature, and is the pre-experiment feature, and is the post-experiment feature. After calculating the variance of the difference regression coefficient of each experiment index, the difference regression coefficient and the variance of the difference regression coefficient can be used as the difference detection information of the experiment index.

[0164] 105. When receiving the difference detection request for the target experiment index, the difference detection information is used to perform difference detection on the target experiment index, and the detection result is sent to the terminal.

[0165] Wherein, the difference detection can also be understood as A / A difference detection and modified A / B difference detection on the target experiment index. Specifically, it can be understood as calculating the pre-experiment (A / A) difference significance probability of the target experiment index, the modified post-experiment (A / B period) difference significance probability and the modified A / B difference information. When the A / A difference significance probability does not exceed the preset threshold, it can be determined that the target experiment index has pre-experiment difference.

[0166] Wherein, when receiving the difference detection request for the target experiment index, the pre-experiment difference detection on the target experiment index can be performed in various ways according to the difference detection information. Specifically, it can be as follows:

[0167] For example, when receiving the difference detection request for the target experiment index sent by the user through the terminal, the target difference detection information corresponding to the target experiment index is screened out in the difference detection information, the difference data of the target experiment index data is calculated according to the target difference detection information, and the detection result of the target experiment index is determined based on the difference data.

[0168] wherein the target difference detection information can be a difference regression coefficient corresponding to the target experiment index and a variance of the difference regression coefficient.

[0169] wherein the difference data can be understood as a pre-experiment (A / A) difference significance probability P AA , a modified post-experiment (A / B) difference significance probability P AB and a modified post-experiment (A / B) difference Δ AB . According to the target difference detection information, there can be multiple ways to calculate the difference data of the target experiment index, such as extracting the target difference regression coefficient and the target variance of the target difference regression coefficient from the target difference detection information, identifying the post-experiment difference regression coefficient from the target difference regression coefficient, and fusing the post-experiment difference regression coefficient with the experiment cumulative days to obtain the modified post-experiment difference information, calculating the difference significance probability of the target experiment index based on the target difference regression coefficient and the target variance, and taking the modified post-experiment difference information and the difference significance probability as the difference data.

[0170] wherein the modified post-experiment difference information can be understood as the difference information after the post-experiment difference information is modified. There can be multiple ways to calculate this modified post-experiment difference information, such as directly multiplying the post-experiment difference regression coefficient by the experiment cumulative days to obtain the modified post-experiment difference information, which can be specifically shown in formula (10):

[0171]

[0172] wherein Δ AB is the modified post-experiment difference information, is the post-experiment difference regression coefficient, and ab_days is the experiment cumulative days, which can be understood as the number of days that the experiment index is observed cumulatively during the experiment.

[0173] wherein the difference significance probability can be understood as probability information indicating whether there is a significant difference in the experiment index during the experiment. The difference significance probability can include a pre-experiment difference significance probability P AA and a modified post-experiment difference significance probability P ABThere are several ways to calculate the probability of significant difference. For example, the target difference regression coefficient can be fused with the target variance to obtain the target detection statistic of the target experimental indicator. The basic probability information that the detection statistic does not exceed the target detection statistic can be obtained. Based on the basic probability information, the target probability information that the detection statistic exceeds the target detection statistic can be calculated. Based on the target probability information, the probability of significant difference of the target experimental indicator can be determined. The probability of significant difference before the experiment and the corrected probability of significant difference after the experiment can be extracted from the probability of significant difference as the probability of significant difference of the target experimental indicator.

[0174] The target detection statistic can be understood as the T-statistic of the target experimental index. The target detection statistic can be obtained by fusing the target difference regression coefficient with the target variance. There are several ways to fuse the target difference regression coefficient with the target variance. For example, the ratio between the target difference regression coefficient and the target variance can be calculated to obtain the target detection statistic of the target experimental index, as shown in formula (11).

[0175]

[0176] Among them, T K For target detection statistics, The target difference regression coefficient, The target variance is the regression coefficient of the target difference.

[0177] Among them, the probability information of difference significance can be understood as information used to indicate the probability of difference significance of the target experimental index. Based on the target probability information, there are many ways to determine the probability information of difference significance of the target experimental index. For example, the target probability information can be multiplied by a preset coefficient to obtain the probability information of difference significance, as shown in formula (12):

[0178] P K =2*(1-P(T≤|T)) K |)) (12)

[0179] Among them, P K For the probability information of the significance of the difference, T K For the target statistic, P(T≤|T K |) is based on probability information. Because... Then the corresponding P K It can include P 0k P 1k P 2k and P 3k The probability of significant difference before the experiment, P. AA =P 2k The corrected post-experimental significance probability PAB = P 3k .

[0180] The pre-experiment difference significance probability P AA , the corrected post-experiment difference significance probability P AB and the corrected post-experiment difference information Δ AB are taken as the difference data of the target experiment index. After the difference data of the target experiment index is calculated, the detection result of the target experiment index can be determined. There are various ways to determine the detection result, for example, the pre-experiment difference significance probability is extracted from the difference data, when the pre-experiment difference significance probability does not exceed the preset probability threshold, the difference data is taken as the detection result of the target experiment index, and when the pre-experiment difference significance probability exceeds the preset probability threshold, the post-experiment experiment index data corresponding to the target experiment index is subjected to difference detection to obtain the detection result.

[0181] The preset probability threshold can be set according to actual application, for example, it can be 0.05 or other values, when the pre-experiment (A / A) difference significance probability does not exceed the preset probability threshold, it means that the A / A difference of the target experiment index is significant, then the difference data of the DID, that is, the pre-experiment difference significance probability P AA , the corrected post-experiment difference significance probability P AB and the corrected post-experiment difference information Δ AB are taken as the detection result. When the pre-experiment (A / A) difference significance probability exceeds the preset probability threshold, it means that the target experiment index does not exist A / A difference, at this time, the post-experiment experiment index data corresponding to the target experiment index can be subjected to difference detection, and the difference detection method can be T test, and the result of the T test is taken as the detection result of the target experiment index. The specific detection process can be shown as in Figure 6 .

[0182] After the detection result of the target experiment index is determined, the experiment result can be sent to the terminal, so that the terminal displays the experiment data processing page. There are various ways to send the experiment result to the terminal, for example, the detection result of the target experiment index can be directly sent to the terminal, or when the number of target experiment indexes to be detected is large, the terminal can also be sent prompt information of the detection result, and the terminal obtains the detection result of the target experiment index in the first experiment data processing device based on the prompt information.

[0183] Among them, the biggest difference between the present scheme and the existing DID difference detection method is that the individual difference parameter is introduced in the present scheme, and the correlation between the pre-experiment and post-experiment data is fully considered. The present scheme is Fast-DID, and the detection efficiency between Fast-DID and DID is evaluated through experiments. The evaluation method can be as follows:

[0184] The use duration of the experimental subjects is selected as the observation index. The experimental subjects are randomly divided into 50 groups for 7 days before and after the experiment. 25 groups are randomly selected as the experimental group, and the other 25 groups are selected as the control group. Repeat simulation 100 times, each time randomly select 2% of the data, and each data has about 50% of the experimental subjects in the experimental group and the other 50% of the experimental subjects in the control group. Observe the distribution of A / A difference significance P value, the modified A / B difference significance P value, and the difference between Fast-DID and general DID in the standard error of the regression coefficient estimate.

[0185] (1) Simulation 1: A / A difference is 0, and the modified A / B difference is 0

[0186] Comparison result 1: The type I error rate of Fast-DID in testing the two differences is about 5%, and the P value approximately obeys the uniform distribution [0, 1]. The type I error rate of general DID in testing the two differences is also about 5%, but the P value of testing the modified A / B difference is larger. The specific comparison results can be shown as Figure 7

[0187] (2) Simulation 2: A / A difference is 2%, and the modified A / B difference is 0

[0188] Comparison result 2: The power of Fast-DID in testing A / A difference is close to 100%, the type I error rate of testing the modified A / B difference is less than 5%, and the P value approximately obeys the uniform distribution [0, 1]. The power of general DID in testing A / A difference is close to 100%, the type I error rate of testing the modified A / B difference is less than 5%, but the P value is larger. The specific comparison results can be shown as Figure 8

[0189] (3) Simulation 3: A / A difference is 0, and the modified A / B difference is 2%

[0190] Comparison result 3: The type I error rate of Fast-DID in testing A / A difference is less than 5%, and the P value approximately obeys the uniform distribution [0, 1], and the power of testing the modified A / B difference is close to 100%. The type I error rate of general DID in testing A / A difference is less than 5%, and the P value approximately obeys the uniform distribution [0, 1], and the power of testing the modified A / B difference is about 80%, which is lower than Fast-DID. The specific comparison results can be shown as Figure 9

[0191] (4) A / A difference is 2%, and the modified A / B difference is 2%

[0192] ​​​Comparison Result 4: The Power of Fast-DID to test the two differences is all greater than 80%. The Power of generalized DID to test the two differences is all greater than 80%, and the Power of generalized DID to test the adjusted A / B difference is lower than Fast-DID, which can be seen in detail as follows. Figure 10

[0193] For Fast-DID, the actual business data is used to evaluate the type I error rate and the Power of the test, and the evaluation results show that Fast-DID can control the type I error rate within 5% and the Power above 80%, and the distribution of the statistical quantity output by the model meets the expectation. In addition, compared with the generalized DID, Fast-DID can reduce the estimated standard error by about 50% when testing the significance of the adjusted A / B difference, which greatly improves the test efficiency.

[0194] The background of Fast-DID is that the experimental effect needs to be adjusted when there is an A / A difference in the experimental scenario. If no adjustment is made, the consequences may include:

[0195] (1) The type I error rate increases. When there is an A / A difference but no real experimental effect, only testing the post-experiment data may give a significant conclusion, which increases the type I error rate.

[0196] (2) The type II error rate increases. When there is an A / A difference and a real experimental effect, but the direction of the A / A difference is opposite to that of the real experimental effect, only testing the post-experiment data may give a non-significant conclusion, which increases the type II error rate.

[0197] Therefore, adding the judgment and adjustment of the A / A difference in the experimental scenario is beneficial to reduce the type I error rate and the type II error rate.

[0198] ​From the above, after obtaining the experimental index data set, the embodiment of the application constructs the to-be-processed experimental data set of the experimental index according to the experimental grouping and data type of the experimental index data in the experimental index data set, then identifies the difference regression coefficient of each experimental index in the to-be-processed experimental data set based on the preset individual difference parameter, filters out the panel experimental data in the to-be-processed experimental data set, and calculates the variance of the difference regression coefficient based on the panel experimental data to obtain the difference detection information of each experimental index, and then when receiving the difference detection request for the target experimental index, difference detection is performed on the target experimental index according to the difference detection information, and the detection result is sent to the terminal. Since the preset difference parameter is introduced in the scheme, the correlation of the sample data before and after the experiment is fully considered, not only the panel data and the mixed cross-section data can be processed, but also the mixed data composed of the panel data and the mixed cross-section data can be processed, the different experimental indexes of the same experiment can be detected and processed at the same time, thereby reducing the consumption of computing resources, and therefore the processing efficiency of the experimental data processing can be improved.

[0199] The embodiment will be described from the perspective of the second experimental data processing apparatus, which can be integrated in an electronic device, which can be a terminal or the like. The terminal can include a tablet computer, a notebook computer, a personal computer (PC), a wearable device, a virtual reality device, or other smart devices that can process experimental data.

[0200] An experimental data processing method includes:

[0201] Sending a difference detection request for a target experimental index to a server, receiving a detection result of the target experimental index returned by the server, displaying an experimental data processing page, the experimental data processing page including a difference significance display control corresponding to the type of the detection result, and displaying the detection result of the target experimental index in response to a display operation on the difference significance display control.

[0202] As shown in Figure 11 , the specific process of the experimental data processing method can be as follows:

[0203] 201. Send a difference detection request for a target experimental index to a server.

[0204] For example, an experiment index selection page including an experiment index list can be displayed, in response to a selection operation on the experiment index list, an experiment index corresponding to the selection operation is selected as a target experiment index, a difference detection request corresponding to the target experiment index is generated based on the target experiment index, and the difference detection request is sent to the server, so that the server filters out target difference detection information corresponding to the target experiment index in the difference detection information, calculates difference data of the target experiment index according to the target difference detection information, and determines a detection result of the target experiment index based on the difference data.

[0205] 202. Receive the detection result of the target experiment index returned by the server.

[0206] For example, the detection result of the target experiment index returned by the server can be directly received, or the target experiment index returned by the server can also be received. Prompt information of the target experiment index is extracted from the prompt information, and the storage address of the detection result is extracted from the prompt information. Based on the storage address, the detection result of the target experiment index is obtained in the first experiment data processing device.

[0207] 203. Display an experiment data processing page.

[0208] The experiment data processing page includes a difference significance display control corresponding to the type of the detection result.

[0209] The type of the detection result can be various, for example, it can include two types of pre-experiment (A / A) difference and no pre-experiment (A / A) difference.

[0210] The experiment data processing page can be displayed in various ways, which can be as follows:

[0211] For example, when the type of the detection result is the pre-experiment difference, the corrected post-experiment difference significance probability is extracted from the detection result, and the experiment data processing page is displayed based on the corrected post-experiment difference significance probability. When the type of the detection result is no pre-experiment difference, the difference significance display control corresponding to the post-experiment difference is obtained, and the experiment data processing page is displayed based on the difference significance display control corresponding to the post-experiment difference.

[0212] The display manner of the experiment data processing page based on the corrected experiment post-difference significance probability can be various. For example, when the corrected experiment post-difference significance probability does not exceed the preset significance probability threshold, it is determined that the correction type of the target experiment index is a correction significant type, and a difference significance display control corresponding to the correction significant type is acquired to display the experiment data processing page. When the corrected experiment post-difference significance probability exceeds the preset significance probability threshold, it is determined that the correction type of the target experiment index is a correction non-significant type, and a difference significance display control corresponding to the correction non-significant type is acquired to display the experiment data processing page.

[0213] The type of the difference significance display control found in the experiment data processing page can be three types. Taking the experiment data processing page as an example, the three types of the difference significance display control are as follows. Figure 12 For example, the experiment index A has an A / A difference, and the correction type is a correction significant type. The difference significance display control can be a full battery cell, and the outer edge is blue. The experiment index B has an A / A difference, and the correction type is a correction non-significant type. The difference significance display control is an empty battery cell, and the outer edge is blue. The experiment index C does not have an A / A difference, and does not need to be corrected. The difference significance display control can be a full battery cell, and the outer edge is gray.

[0214] 204. In response to a display operation on the difference significance display control, the detection result of the target experiment index is displayed.

[0215] For example, in response to a display operation on the difference significance display control, the display content is determined according to the detection result of the target experiment index, and the display content is displayed on the experiment data processing page.

[0216] The manner of determining the display content according to the detection result of the target experiment index can be various. For example, when the detection result is that the target experiment index has a pre-experiment difference, and the correction type is a correction significant type, the display content can be the cumulative mean of the experimental group, the cumulative mean of the control group, the cumulative absolute difference, the cumulative correction absolute value, the cumulative correction difference ratio, and the detection result (P AA , P AB , and Δ AB ). When the detection result is that the target experiment index has a pre-experiment difference, and the correction type is a correction non-significant type, the display content is the same as that when the correction type is a correction non-significant type. When the detection result is that there is no pre-experiment difference, the display content can be the cumulative mean of the experimental group, the cumulative mean of the control group, the cumulative absolute difference, the cumulative relative difference ratio, and the T-test significance (T-test result).

[0217] From the above, the embodiment of the application sends a difference detection request for the target experimental index to the server, receives the detection result of the target experimental index returned by the server, and then displays an experimental data processing page. The experimental data processing page includes a difference significance display control corresponding to the type of the detection result. In response to a display operation on the difference significance display control, the detection result of the target experimental index is displayed. Since the scheme can indicate the detection result of the target experimental index through the type of the difference display control, the detection result can be quickly displayed. Moreover, the detection result of the target experimental index can be triggered to be displayed through the display operation. Therefore, the processing efficiency of experimental data processing can be improved.

[0218] According to the method described in the above embodiment, the following will be further described by way of example.

[0219] In this embodiment, the first experimental data processing device is specifically integrated in the first electronic device, the first electronic device is a server, the second experimental data processing device is specifically integrated in the second electronic device, and the second electronic device is a terminal.

[0220] As shown in Figure 13 , an experimental data processing method includes the following specific process:

[0221] 301. The server obtains an experimental index data set.

[0222] For example, the server can obtain at least one current experimental index data of at least one experimental object on the current experimental day, obtain experimental index cumulative data of each experimental index, query target experimental index cumulative data corresponding to each current experimental index data in the experimental index cumulative data according to the experimental index corresponding to the current experimental index data, add the current experimental index data to the target experimental index cumulative data when the target experimental index cumulative data exists, obtain updated experimental index cumulative data, and add the current experimental index data as updated experimental index cumulative data when the target experimental index cumulative data does not exist. Alternatively, the experimental index data of multiple days can be directly added to obtain updated experimental index cumulative data. The experimental cumulative days corresponding to each updated experimental index cumulative data of the experimental object are obtained, and the ratio between each updated experimental index cumulative data and the corresponding experimental cumulative days is calculated to obtain the experimental index data of each experimental index of the experimental object. The experimental index data of each experimental object is combined to obtain an experimental index data set.

[0223] 302. The server constructs a to-be-processed experimental data set for the experimental index according to the experimental grouping and data type of the experimental index data.

[0224] For example, the server can obtain attribute information of each experimental index data in the experimental index data set, identify the experimental group and data type of the experimental index data in the attribute information, when the experimental group is the experimental group, the corresponding experimental group parameter A can be 1, when the experimental group is the control group, the corresponding experimental group parameter A can be 0. When the data type is pre-experiment data, the corresponding data type parameter T can be 0, when the data type is post-experiment data, the corresponding data type parameter T can be 1. The experimental group parameter and the data type parameter are fused to obtain the experimental index fusion parameter, and the experimental index fusion parameter, the experimental group parameter, the data type parameter and the experimental index data are combined to obtain the to-be-processed experimental data set.

[0225] 303、The server identifies the difference regression coefficient of each experimental index in the to-be-processed experimental data set based on the preset individual difference parameter.

[0226] For example, the server can extract mixed cross-sectional data and panel data from the to-be-processed experimental data set according to the data type and the experimental group, divide the mixed cross-sectional data into pre-experiment cross-sectional data and post-experiment cross-sectional data according to the experimental group, divide the panel data into pre-experiment panel data and post-experiment panel data according to the experimental group, take the pre-experiment panel data and the pre-experiment cross-sectional data as pre-experiment sample data, and take the post-experiment panel data and the post-experiment cross-sectional data as post-experiment sample data. Therefore, the classification of the experimental index data in the to-be-processed experimental data set can be divided into panel data, pre-experiment sample data and post-experiment sample data. The panel data sample size, A / A period sample size and A / B period sample size can also be aggregated, and the sum, square sum and product of the pre-experiment index data and the post-experiment index data of the experimental index data can also be aggregated. The intermediate result obtained by aggregation is used as the initial intermediate statistic of the experimental index. The initial intermediate statistic is fused to obtain the intermediate statistic of the experimental index.

[0227] The server regards the data area corresponding to the data row with null data in the to-be-processed experimental data set df as the data area of the experimental index data area, directly deletes other data in the data area, or can also directly delete the data area in the to-be-processed experimental data set df, to obtain the target experimental data set df filt In the target experimental data set df filt identify the number of rows N filt , each row is regarded as a data area, therefore, the number of rows N filt in df filt can be the number of regions.

[0228] The server extracts experiment parameters in the target experiment data set to obtain an experiment parameter set, extracts experiment index data in the target experiment data set to obtain a target experiment index data set, extracts features from the experiment parameter set to obtain experiment parameter features, and extracts features from the target experiment index data set to obtain experiment index features.

[0229] The server constructs a first basic feature vector X filt =(1,T,A,T·A), and takes experiment parameters in each row of the experiment parameter set as elements in the first basic feature vector, so as to obtain a first feature vector corresponding to the experiment parameters, and take the first feature vector as the experiment parameter features. A second basic feature vector Y filt =(Y1,.....,Y K ) is constructed, and experiment index data of each experiment index in the target experiment index data set is taken as an element in the second basic feature vector, so as to obtain a second feature vector corresponding to the experiment index, and take the second feature vector as the experiment index features.

[0230] The server transposes the first feature vector X filt to obtain a transposed first feature vector The transposed first feature vector is taken as the converted experiment parameter features. Based on a preset individual difference parameter and N filt , the transposed first feature vector is multiplied by the first feature vector X filt and the second basic feature vector Y filt , so as to obtain fused experiment parameter features, which can be specifically shown in formula (1).

[0231] The server multiplies the transposed first feature vector and the second basic feature vector Y filt based on a preset individual difference parameter, so as to obtain fused experiment index features, which can be specifically shown in formula (2). The inverse matrix of the fused experiment parameter features is multiplied by the fused experiment index features to obtain a difference regression matrix. Based on the intermediate statistics, an element value corresponding to each element in the difference regression matrix is calculated, the element value is replaced in the corresponding element in the difference regression matrix, so as to obtain difference regression information, which can be specifically shown in formula (3). Each column in the difference regression information represents a DID regression coefficient corresponding to each experiment index, and each column element is taken as a difference regression coefficient of the experiment index, which can be shown in formula (3).

[0232] 304、The server screens out panel data in the to-be-processed experiment data set.

[0233] For example, taking the data set df shown in Table 1 as an example, the server can find in Table 1 that the experimental objects c and d only have pre-experiment experimental index data, and the experimental objects e and f only have post-experiment experimental index data, and thus the data in the rows of the experimental objects c, d, e, and f can be filtered out, so that the panel data df both It can be found from Table 1 that the panel data can be the experimental index data and experimental parameters of the experimental objects a and b.

[0234] 305、The server calculates the variance of the difference regression coefficient based on the panel data, to obtain difference detection information of each experimental index.

[0235] For example, the server filters out the experimental index data of each experimental object when the data type parameter T is 0 in the panel data to obtain pre-experiment experimental index data, and filters out the experimental index data when the data type T is 1 in the panel data to obtain post-experiment experimental index data. After filtering out the pre-experiment experimental index data and the post-experiment index data, the panel data df both The data format of the panel data is use id ,A,Y 11 ,Y 12 ,......,Y K1 ,Y K2 。Y K1 represents pre-experiment experimental index data, and Y K2 represents post-experiment experimental index data.

[0236] The server obtains preset pre-experiment parameter information X pre =(1,0,A,0) and preset post-experiment parameter information X post =(1,1,A,A), fuses the difference regression coefficient with the preset pre-experiment parameter information to obtain fused pre-experiment data, and fuses the difference regression coefficient with the preset post-experiment parameter information to obtain fused post-experiment data, calculates a data difference value between the pre-experiment experimental index data and the fused pre-experiment data, and takes the data difference value as a pre-experiment residual, which can be specifically shown in formula (5). A difference value between the post-experiment experimental index data and the fused post-experiment data is calculated to obtain a post-experiment residual, which can be specifically shown in formula (6). The pre-experiment residual and the post-experiment residual are taken as residual information of the corresponding experimental index. A residual covariance matrix V K =(Cov(e K1 ,e K2 )) and an inverse matrix W K =(Cov(e K1 ,e K2 ))-1 The residual covariance matrix V K and its inverse matrix W K The residual characteristics are taken as experimental indexes.

[0237] Taking the experimental data set to be processed, the data set df in Table 1, as an example, the server takes the data of the data row of T = 0 as the pre-experiment data, and takes the data of the data row of T = 1 as the post-experiment data, that is, the explanatory variables of df can be understood as being divided into pre-experiment and post-experiment two parts, as shown in formula (7).

[0238] The server extracts the basic pre-experiment characteristics from the pre-experiment data, converts the basic pre-experiment characteristics to obtain converted basic pre-experiment characteristics, fuses the converted basic pre-experiment characteristics and the basic pre-experiment characteristics to obtain first fused pre-experiment characteristics, extracts the basic post-experiment characteristics from the post-experiment data, fuses the converted basic pre-experiment characteristics and the basic post-experiment characteristics to obtain second fused pre-experiment characteristics, and fuses the first fused pre-experiment characteristics and the second fused pre-experiment characteristics to obtain pre-experiment characteristics. The basic post-experiment characteristics are converted to obtain converted post-experiment characteristics, the basic post-experiment characteristics and the converted post-experiment characteristics are fused to obtain first fused post-experiment characteristics, the converted post-experiment characteristics and the basic pre-experiment characteristics are fused to obtain second fused post-experiment characteristics, and the first fused post-experiment characteristics and the second fused post-experiment characteristics are fused to obtain post-experiment characteristics.

[0239] The server divides the pre-experiment residual characteristics and the post-experiment residual characteristics from the residual characteristics, which can be as shown in formula (8). The pre-experiment residual characteristics and the pre-experiment characteristics are fused to obtain fused pre-experiment characteristics, and the post-experiment residual characteristics and the post-experiment characteristics are fused to obtain fused post-experiment characteristics. The fused pre-experiment characteristics and the fused post-experiment characteristics are added to obtain target variance characteristics, the target element values of each element in the matrix corresponding to the target variance characteristics are calculated based on the intermediate statistics, the target characteristic value of the target variance characteristics is determined based on the target element values, and the inverse matrix of the target characteristic value is calculated, so as to obtain the variance of the difference regression coefficient of the experimental index, which can be as shown in formula (9). After calculating the variance of the difference regression coefficient of each experimental index, the difference regression coefficient and the variance of the difference regression coefficient can be taken as the difference detection information of the experimental index.

[0240] 306. The terminal sends a difference detection request for the target experimental index to the server.

[0241] For example, the terminal displays an experimental index selection page including an experimental index list, in response to a selection operation on the experimental index list, selects an experimental index corresponding to the selection operation as a target experimental index, generates a difference detection request corresponding to the target experimental index based on the target experimental index, and sends the difference detection request to the server, so that the server filters out target difference detection information corresponding to the target experimental index from the difference detection information, calculates difference data of the target experimental index according to the target difference detection information, and determines a detection result of the target experimental index based on the difference data.

[0242] 307、The server performs difference detection on the target experimental index according to the difference detection data when receiving the difference detection request for the target experimental index.

[0243] For example, when the server receives a difference detection request for a target experimental index sent by a user through a terminal, the server filters out target difference detection information corresponding to the target experimental index from the difference detection information, extracts a target difference regression coefficient and a target variance of the target difference regression coefficient from the target difference detection information, and identifies a post-experiment difference regression coefficient in the target difference regression coefficient. The post-experiment difference regression coefficient is directly multiplied by the cumulative number of experiments to obtain a corrected post-experiment difference information Δ AB , which can be shown in formula (10).

[0244] The server calculates the ratio between the target difference regression coefficient and the target variance of the target difference regression coefficient to obtain a target detection statistic of the target experimental index, which can be shown in formula (11). The basic probability information of the detection statistic not exceeding the target detection statistic is obtained, and the target probability information of the detection statistic exceeding the target detection statistic is calculated based on the basic probability information. The target probability information is multiplied by a preset coefficient to obtain difference significance probability information P K , which can be shown in formula (12). Correspondingly, P K may include P 0K , P 1K , P 2K , and P 3K , P 2K is taken as pre-experiment difference significance probability P AA , P 3K is taken as corrected post-experiment difference significance probability P AB . The pre-experiment difference significance probability and the corrected post-experiment difference significance probability are taken as the difference significance probability of the target experimental index. Finally, the pre-experiment difference significance probability P AA , the corrected post-experiment difference significance probability P AB , and the corrected post-experiment difference information Δ ABThe difference data as the target experimental index.

[0245] The server extracts the pre-experiment difference significance probability P from the difference data. AA When the pre-experiment (A / A) difference significance probability P AA If the pre-experiment (A / A) difference significance probability P AA , the corrected post-experiment difference significance probability P AB and the corrected post-experiment difference information Δ AB as the detection result. When the pre-experiment (A / A) difference significance probability P AA If the pre-experiment (A / A) difference significance probability P

[0246] 308、The terminal receives the detection result of the target experimental index returned by the server.

[0247] For example, the terminal can directly receive the detection result of the target experimental index returned by the server, or can also receive the prompt information of the target experimental index returned by the server, extract the storage address of the detection result in the prompt information, and obtain the detection result of the target experimental index in the first experimental data processing device based on the storage address.

[0248] 309、The terminal displays the experimental data processing page.

[0249] For example, when the type of the detection result is the existence of pre-experiment difference, the terminal extracts the corrected post-experiment difference significance probability in the detection result, determines that the correction type of the target experimental index is the correction significant type when the corrected post-experiment difference significance probability does not exceed the preset significance probability threshold, and obtains the difference significance display control corresponding to the correction significant type to display the experimental data processing page. When the corrected post-experiment difference significance probability exceeds the preset significance probability threshold, it is determined that the correction type of the target experimental index is the correction non-significant type, and the difference significance display control corresponding to the correction non-significant type is obtained to display the experimental data processing page. When the type of the detection result is the non-existence of pre-experiment difference, the terminal obtains the difference significance display control corresponding to the post-experiment difference, and displays the experimental data processing page based on the difference significance display control corresponding to the post-experiment difference.

[0250] The difference significance display control can be a full battery, with a blue outer edge, when experimental indicator A has an A / A difference and the correction type is a significant correction type, a full battery with a blue outer edge when experimental indicator B has an A / A difference and the correction type is a non-significant correction type, and a full battery with a gray outer edge when experimental indicator C has no A / A difference and no correction is needed.

[0251] 310. The terminal displays the detection result of the target experimental indicator in response to the display operation on the difference significance display control.

[0252] For example, the terminal displays the content as the cumulative mean of the experimental group, the cumulative mean of the control group, the cumulative absolute difference, the cumulative correction absolute value, the cumulative correction difference ratio, and the detection result (P AA , P AB , and Δ AB ) when the detection result is that the target experimental indicator has a pre-experiment difference and the correction type is a significant correction type in response to the display operation on the difference significance display control. The content displayed when the detection result is that the target experimental indicator has a pre-experiment difference and the correction type is a non-significant correction type is the same as the content displayed when the correction type is a non-significant correction type. The content displayed when the detection result is that there is no pre-experiment difference is the cumulative mean of the experimental group, the cumulative mean of the control group, the cumulative absolute difference, the cumulative relative difference ratio, and the T-test significance (T-test result). The content is displayed on the experimental data processing page.

[0253] As can be seen from the above, the server in this embodiment, after obtaining the experimental indicator data set, constructs a to-be-processed experimental data set for the experimental indicators according to the experimental grouping and data type of the experimental indicator data in the experimental indicator data set, then identifies the difference regression coefficient of each experimental indicator in the to-be-processed experimental data set based on a preset individual difference parameter, filters panel experimental data in the to-be-processed experimental data set, calculates the variance of the difference regression coefficient based on the panel experimental data, obtains the difference detection information of each experimental indicator, and then, when receiving a difference detection request for a target experimental indicator, performs difference detection on the target experimental indicator according to the difference detection information and sends the detection result to the terminal. Since the preset difference parameter is introduced in this scheme, the correlation between the sample data before and after the experiment is fully considered, not only panel data and mixed cross-sectional data can be processed, but also mixed data composed of panel data and mixed cross-sectional data can be processed, different experimental indicators of the same experiment can be detected and processed at the same time, thereby reducing the consumption of computing resources, and thus the processing efficiency of experimental data processing can be improved.

[0254] To better implement the above method, the embodiment of the present application further provides an experimental data processing device (i.e., a first experimental data processing device), which can be integrated in a server, which can be a single server or a server cluster composed of multiple servers.

[0255] For example, as shown in Figure 14 The first experimental data processing device can include an acquisition unit 401, a construction unit 402, an identification unit 403, a calculation unit 404, and a detection unit 405, as follows:

[0256] (1) The acquisition unit 401;

[0257] The acquisition unit 401 is configured to acquire an experimental index data set, which includes experimental index data of at least one experimental index of at least one experimental object under a current experiment.

[0258] For example, the acquisition unit 401 can be specifically configured to acquire at least one current experimental index data of at least one experimental object under a current experimental day, update the corresponding experimental index cumulative data based on the current experimental index data, and calculate the ratio of each updated experimental index cumulative data to the experimental cumulative day to obtain the experimental index data set.

[0259] (2) The construction unit 402;

[0260] The construction unit 402 is configured to construct a to-be-processed experimental data set for an experimental index according to the experimental grouping and data type of the experimental index data.

[0261] For example, the construction unit 402 can be specifically configured to acquire attribute information of each experimental index data in the experimental index data set, identify the experimental grouping and data type of the experimental index data in the attribute information, determine the experimental grouping parameter of the experimental index data based on the experimental grouping, and determine the data type parameter of the experimental index data based on the data type, and fuse the experimental grouping parameter and the data type parameter to obtain the experimental index fusion parameter, and combine the experimental index fusion parameter, the experimental grouping parameter, the data type parameter, and the experimental index data to obtain the to-be-processed experimental data set.

[0262] (3) The identification unit 403;

[0263] The identification unit 403 is configured to identify a difference regression coefficient of each experimental index in the to-be-processed experimental data set based on a preset individual difference parameter.

[0264] For example, the identifying unit 403 can be specifically configured to aggregate intermediate statistical quantities of the experimental indexes in the to-be-processed experimental data set according to the experimental grouping and the data type, identify data regions with missing experimental index data in the to-be-processed experimental data set, filter data in the data regions in the to-be-processed experimental data set to obtain a target experimental data set, and identify difference regression coefficients of each experimental index in the target experimental data set according to preset individual difference parameters and the intermediate statistical quantities.

[0265] (4) the calculating unit 404;

[0266] The calculating unit 404 is configured to filter panel experimental data in the to-be-processed experimental data set, and calculate variance of the difference regression coefficients based on the panel experimental data to obtain difference detection data of each experimental index.

[0267] For example, the calculating unit 404 can be specifically configured to filter data of experimental subjects with only one observation in the to-be-processed experimental data set to obtain panel data. According to the difference regression coefficients, residual features of corresponding experimental indexes are extracted from the panel data, pre-experimental data and post-experimental data are segmented from the to-be-processed experimental data set, variance of the difference regression coefficients of corresponding experimental index data is calculated based on the intermediate statistical quantities, the residual features, the pre-experimental data and the post-experimental data, and the difference regression coefficients and the variance of the difference regression coefficients are taken as the difference detection data of the experimental indexes.

[0268] (5) the detecting unit 405;

[0269] The detecting unit 405 is configured to perform difference detection on a target experimental index according to the difference detection data when receiving a difference detection request for the target experimental index, and send a detection result to a terminal.

[0270] For example, the detecting unit 405 can be specifically configured to filter target difference detection data corresponding to the target experimental index from the difference detection data when receiving a difference detection request for the target experimental index sent by a user through the terminal, calculate difference data of the target experimental index data according to the target difference detection data, and determine a detection result of the target experimental index based on the difference data.

[0271] In specific implementation, the above units can be implemented as independent entities, or can be combined as the same or several entities, and the specific implementation of the above units can be referred to the method embodiments above, which will not be described here.

[0272] From the above, after the acquisition unit 401 acquires the experimental index data set, the construction unit 402 constructs the to-be-processed experimental data set for the experimental index according to the experimental grouping and the data type of the experimental index data in the experimental index data set, then the identification unit 403 identifies the difference regression coefficient of each experimental index in the to-be-processed experimental data set based on the preset individual difference parameter, the calculation unit 404 filters out the panel experimental data in the to-be-processed experimental data set, and calculates the variance of the difference regression coefficient based on the panel experimental data, obtains the difference detection data of each experimental index, then the detection unit 405 performs difference detection on the target experimental index according to the difference detection data when receiving the difference detection request for the target experimental index, and sends the detection result to the terminal; Since the preset difference parameter is introduced in the scheme, the correlation of the sample data before and after the experiment is fully considered, not only the panel data and the mixed cross-section data can be processed, but also the mixed data composed of the panel data and the mixed cross-section data can be processed, the different experimental indexes of the same experiment can be detected and processed at the same time, thereby reducing the consumption of computing resources, and therefore, the processing efficiency of experimental data processing can be improved.

[0273] In order to better implement the above method, the embodiment of the application also provides an experimental data processing device (namely a second experimental data processing device), which can be integrated in a terminal. The terminal can include a tablet computer, a notebook computer, and / or a personal computer, etc.

[0274] For example, as shown in Figure 15 The second experimental data processing device can include a sending unit 501, a receiving unit 502, a first display unit 503, and a second display unit 504, as follows:

[0275] (1) The sending unit 501;

[0276] The sending unit 501 is configured to send a difference detection request for a target experimental index to a server.

[0277] For example, the sending unit 501 can be specifically configured to display an experimental index selection page, the experimental index selection page including an experimental index list, in response to a selection operation on the experimental index list, taking the experimental index corresponding to the selection operation as a target experimental index, generating a difference detection request corresponding to the target experimental index based on the target experimental index, and sending the difference detection request to the server, so that the server filters out target difference detection data corresponding to the target experimental index in the difference detection data, calculates difference data of the target experimental index according to the target difference detection data, and determines a detection result of the target experimental index based on the difference data.

[0278] (2) The receiving unit 502;

[0279] The receiving unit 502 is configured to receive the detection result of the target experiment index returned by the server.

[0280] For example, the receiving unit 502 can be specifically configured to receive the detection result of the target experiment index returned by the server, or can also receive prompt information of the target experiment index returned by the server, extract a storage address of the detection result in the prompt information, and acquire the detection result of the target experiment index in the first experiment data processing device based on the storage address.

[0281] (3) The first display unit 503;

[0282] The first display unit 503 is configured to display an experiment data processing page, and the experiment data processing page includes a difference significance display control corresponding to the type of the detection result.

[0283] For example, the first display unit 503 can be specifically configured to extract a corrected post-experiment difference significance probability in the detection result when the type of the detection result is the pre-experiment difference, and display the experiment data processing page based on the corrected post-experiment difference significance probability, and when the type of the detection result is the pre-experiment difference, acquire a difference significance display control corresponding to the post-experiment difference, and display the experiment data processing page based on the difference significance display control corresponding to the post-experiment difference.

[0284] (4) The second display unit 504;

[0285] The second display unit 504 is configured to display the detection result of the target experiment index in response to a display operation on the difference significance display control.

[0286] For example, the second display unit 504 can be specifically configured to determine display content according to the detection result of the target experiment index in response to the display operation on the difference significance display control, and display the display content on the experiment data processing page.

[0287] In specific implementation, each of the above units can be implemented as an independent entity, or can be combined as the same or several entities, and the specific implementation of each of the above units can be referred to the method embodiments above, which will not be described here.

[0288] From the above, the embodiment of the present application sends the difference detection request for the target experiment index to the server by the sending unit 501, receives the detection result of the target experiment index returned by the server by the receiving unit 502, then the first display unit 503 displays the experiment data processing page, the experiment data processing page includes the difference significance display control corresponding to the type of the detection result, and the second display unit 504 displays the detection result of the target experiment index in response to the display operation of the difference significance display control. Since the scheme can indicate the detection result of the target experiment index through the type of the difference display control, the detection result can be quickly displayed, and the display of the detection result of the target experiment index can be triggered through the display operation, so that the processing efficiency of the experiment data processing can be improved.

[0289] The embodiment of the present application also provides an electronic device, as shown in the figure, which shows a structural schematic diagram of the electronic device related to the embodiment of the present application, in particular: Figure 16

[0290] The electronic device can include a processor 601 with one or more processing cores, a memory 602 with one or more computer readable storage media, a power supply 603, an input unit 604 and the like. Those skilled in the art can understand that the structure of the electronic device shown in the figure does not constitute a limitation on the electronic device, and can include more or fewer components than the figure, or combine certain components, or different component arrangements. Among them: Figure 16

[0291] The processor 601 is the control center of the electronic device, which connects all parts of the electronic device through various interfaces and lines, and performs various functions and processes data of the electronic device by running or executing software programs and / or modules stored in the memory 602 and calling data stored in the memory 602. Optionally, the processor 601 can include one or more processing cores; preferably, the processor 601 can integrate an application processor and a modem processor, wherein the application processor mainly processes the operating system, user interface and application programs, and the modem processor mainly processes wireless communication. It can be understood that the above-mentioned modem processor can also not be integrated into the processor 601.

[0292] ​​The memory 602 can be used to store software programs and modules, and the processor 601 executes various function applications and data processing by running the software programs and modules stored in the memory 602. The memory 602 can mainly include a program storage area and a data storage area, wherein the program storage area can store an operating system, application programs required by at least one function (such as a sound playing function, an image playing function, etc.), etc.; and the data storage area can store data created according to the use of the electronic device, etc. In addition, the memory 602 can include a high-speed random access memory, and can also include a non-volatile memory, such as at least one magnetic disk storage device, a flash memory device, or other volatile solid-state memory device. Accordingly, the memory 602 can also include a memory controller to provide access for the processor 601 to the memory 602.

[0293] The electronic device also includes a power supply 603 for powering the various components. Preferably, the power supply 603 can be logically connected to the processor 601 through a power management system, so as to realize functions such as management of charging, discharging, and power consumption management through the power management system. The power supply 603 can also include one or more direct current or alternating current power supplies, a recharging system, a power supply fault detection circuit, a power supply converter or inverter, a power supply status indicator, etc.

[0294] The electronic device can also include an input unit 604, which can be used to receive input digital or character information, and generate keyboard, mouse, joystick, optical or trackball signal inputs related to user settings and function controls.

[0295] Although not shown, the electronic device can also include a display unit, etc., which will not be described here. Specifically, in the present embodiment, the processor 601 in the electronic device will load the executable file corresponding to the process of one or more application programs into the memory 602 according to the following instructions, and run the application programs stored in the memory 602 by the processor 601, so as to realize various functions, as follows:

[0296] An experimental index data set is obtained, the experimental index data set including experimental index data of at least one experimental index of at least one experimental object under a current experiment, a to-be-processed experimental data set for the experimental index is constructed according to an experimental grouping and a data type of the experimental index data, a difference regression coefficient of each experimental index is identified in the to-be-processed experimental data set based on a preset individual difference parameter, panel experimental data is screened out in the to-be-processed experimental data set, and a variance of the difference regression coefficient is calculated based on the panel experimental data, to obtain difference detection data of each experimental index, when a difference detection request for a target experimental index is received, difference detection is performed on the target experimental index according to the difference detection data, and a detection result is sent to a terminal.

[0297] or

[0298] sending a difference detection request for a target experiment index to the server, receiving a detection result of the target experiment index returned by the server, and displaying an experiment data processing page, the experiment data processing page including a difference significance display control corresponding to a type of the detection result, and in response to a display operation on the difference significance display control, displaying the detection result of the target experiment index.

[0299] The specific implementation of each operation can refer to the foregoing embodiments, and will not be described here.

[0300] As can be seen from the above, after obtaining the experiment index data set, the embodiment of the application constructs a to-be-processed experiment data set for the experiment index according to the experiment grouping and data type of the experiment index data in the experiment index data set, then identifies the difference regression coefficient of each experiment index in the to-be-processed experiment data set based on a preset individual difference parameter, filters out panel experiment data in the to-be-processed experiment data set, and calculates the variance of the difference regression coefficient based on the panel experiment data to obtain the difference detection data of each experiment index, then when receiving a difference detection request for a target experiment index, the target experiment index is detected according to the difference detection data, and the detection result is sent to the terminal; since the preset difference parameter is introduced in the scheme, the correlation of the sample data before and after the experiment is fully considered, not only the panel data and the mixed cross-section data can be processed, but also the mixed data composed of the panel data and the mixed cross-section data can be processed, the different experiment indexes of the same experiment can be detected at the same time, thereby reducing the consumption of computing resources, and therefore the processing efficiency of the experiment data processing can be improved.

[0301] Those skilled in the art can understand that all or part of the steps in the various methods of the above embodiments can be completed by instructions, or by related hardware controlled by the instructions, the instructions can be stored in a computer readable storage medium and loaded and executed by a processor.

[0302] To this end, the embodiment of the application provides a computer readable storage medium, which stores a plurality of instructions, the instructions can be loaded by a processor to execute the steps in any experiment data processing method provided by the embodiment of the application. For example, the instructions can execute the following steps:

[0303] Obtaining an experimental index data set, the experimental index data set including experimental index data of at least one experimental index of at least one experimental object under a current experiment, constructing a to-be-processed experimental data set for the experimental index according to an experimental grouping and a data type of the experimental index data, identifying a difference regression coefficient of each experimental index in the to-be-processed experimental data set based on a preset individual difference parameter, filtering out panel experimental data in the to-be-processed experimental data set, and calculating a variance of the difference regression coefficient based on the panel experimental data to obtain difference detection data of each experimental index, when a difference detection request for a target experimental index is received, performing difference detection on the target experimental index according to the difference detection data, and sending a detection result to a terminal.

[0304] Or

[0305] Sending a difference detection request for a target experimental index to a server, receiving a detection result of the target experimental index returned by the server, displaying an experimental data processing page, the experimental data processing page including a difference significance display control corresponding to a type of the detection result, and displaying the detection result of the target experimental index in response to a display operation on the difference significance display control.

[0306] The specific implementation of each operation can be referred to the foregoing embodiments, which will not be described here.

[0307] The computer readable storage medium can include a read only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, etc.

[0308] Since the computer readable storage medium stores instructions, the steps of any experimental data processing method provided by the embodiments of the present application can be executed, and thus the beneficial effects of any experimental data processing method provided by the embodiments of the present application can be achieved. Details are described in the foregoing embodiments, which will not be described here.

[0309] According to an aspect of the present application, a computer program product or a computer program is provided, which includes computer instructions stored in a computer readable storage medium. The processor of the computer device reads the computer instructions from the computer readable storage medium, and the processor executes the computer instructions, so that the computer device executes the method provided in the various optional implementation manners of the experimental data processing aspect or the experimental index difference detection aspect.

[0310] The experimental data processing method and device provided by the embodiment of the present application are described in detail above, and the principle and implementation manner of the present application are described by applying specific examples in this paper. The above embodiment description is only used to help understand the method of the present application and its core idea; meanwhile, for those skilled in the art, according to the idea of the present application, the specific implementation manner and application range will be changed, and the above description should not be understood as a limitation on the present application.

Claims

1. A method of processing experimental data, characterized by, The method comprises the following steps: acquiring an experimental index data set, the experimental index data set comprising experimental index data of at least one experimental index of at least one experimental object under a current experiment; constructing a to-be-processed experimental data set for the experimental index according to an experimental grouping and a data type of the experimental index data; identifying a difference regression coefficient of each experimental index in the to-be-processed experimental data set based on a preset individual difference parameter; screening panel data in the to-be-processed experimental data set, and calculating a variance of the difference regression coefficient based on the panel data to obtain difference detection information of each experimental index; when a difference detection request for a target experimental index is received, performing difference detection on the target experimental index according to the difference detection information, and sending a detection result to a terminal.

2. The experimental data processing method of claim 1, wherein, The method comprises the following steps: aggregating an intermediate statistic of the experimental index in the to-be-processed experimental data set according to the experimental grouping and the data type; identifying a data region with missing experimental index data in the to-be-processed experimental data set; filtering data in the data region to obtain a target experimental data set; identifying a difference regression coefficient of each experimental index in the target experimental data set according to the preset individual difference parameter and the intermediate statistic.

3. The experimental data processing method of claim 2, wherein, The method comprises the following steps: acquiring a region quantity of the data region in the target experimental data set; extracting an experimental index feature and an experimental parameter feature in the target experimental data set; fusing the experimental index feature and the experimental parameter feature based on the intermediate statistic, the preset individual difference parameter and the region quantity to obtain the difference regression coefficient of each experimental index.

4. The experimental data processing method of claim 3, wherein, The method comprises the following steps: performing feature conversion on the experimental parameter feature to obtain a converted experimental parameter feature; fusing the converted experimental parameter feature and the experimental parameter feature based on the region quantity and the preset individual difference parameter to obtain a fused experimental parameter feature; fusing the converted experimental parameter feature and the experimental index feature based on the preset individual difference parameter to obtain a fused experimental index feature; determining the difference regression coefficient of each experimental index based on the intermediate statistic, the fused experimental parameter feature and the fused experimental index feature.

5. The experimental data processing method of claim 4, wherein, The method comprises the following steps: fusing the fused experimental parameter feature and the fused experimental index feature based on the intermediate statistic to obtain difference regression information; Extracting a difference regression coefficient of each experimental index in the difference regression information.

6. The experimental data processing method of claim 2, wherein, The intermediate statistical quantity of the experimental index is aggregated in the to-be-processed experimental data set according to the data type, and the intermediate statistical quantity of the experimental index is aggregated in the to-be-processed experimental data set according to the data type, including: According to the experimental grouping and the data type, the experimental index data in the to-be-processed experimental data set is classified; The initial intermediate statistical quantity of the experimental index is aggregated in each type of experimental index data respectively; The initial intermediate statistical quantity is fused to obtain the intermediate statistical quantity of the experimental index.

7. The experimental data processing method according to any one of claim 2, characterized in that, The variance of the difference regression coefficient is calculated based on the panel data to obtain the difference detection information of each experimental index, including: According to the difference regression coefficient, the residual feature of the corresponding experimental index is extracted in the panel data; The pre-experiment data and the post-experiment data are cut out from the to-be-processed experimental data set; Based on the intermediate statistical quantity, the residual feature, the pre-experiment data and the post-experiment data, the variance of the difference regression coefficient of the corresponding experimental index is calculated; The difference regression coefficient and the variance of the difference regression coefficient are taken as the difference detection information of the experimental index.

8. The experimental data processing method of claim 7, wherein, According to the difference regression coefficient, the residual feature of the corresponding experimental index is extracted in the panel data, including: The pre-experiment index data and the post-experiment index data corresponding to each experimental index are screened out in the panel data; Based on the difference regression coefficient, the pre-experiment index data and the post-experiment index data, the residual information of the corresponding experimental index is determined; The residual feature of the experimental index is extracted in the residual information.

9. The experimental data processing method of claim 6, wherein, Based on the difference regression coefficient, the pre-experiment index data and the post-experiment index data, the residual information of the corresponding experimental index is determined, including: Obtaining preset pre-experiment parameter information and preset post-experiment parameter information; The difference regression coefficient is fused with the preset pre-experiment parameter information to obtain fused pre-experiment data, and the difference regression coefficient is fused with the preset post-experiment parameter information to obtain fused post-experiment data; The residual of the pre-experiment index data and the fused pre-experiment data is calculated to obtain the pre-experiment residual, and the residual of the post-experiment index data and the fused post-experiment data is calculated to obtain the post-experiment residual; The pre-experiment residual and the post-experiment residual are taken as the residual information of the corresponding experimental index.

10. The experimental data processing method of claim 7, wherein, Based on the intermediate statistical quantity, the residual feature, the pre-experiment data and the post-experiment data, the variance of the difference regression coefficient of the corresponding experimental index is calculated to obtain the coefficient variance, including: The pre-experiment feature is extracted in the pre-experiment data, and the post-experiment feature is extracted in the post-experiment data; The residual feature is fused with the pre-experiment feature to obtain the fused pre-experiment feature, and the residual feature is fused with the post-experiment feature to obtain the fused post-experiment feature; Based on the intermediate statistical quantity, the fused pre-experiment feature and the fused post-experiment feature are fused to obtain the variance of the difference regression coefficient of the corresponding experimental index.

11. The experimental data processing method according to any one of claims 1 to 10, characterized in that, The difference detection information is used for detecting the target experiment index, and the detecting comprises the following steps. The target difference detection information corresponding to the target experiment index is screened out from the difference detection information. Difference data of the target experiment index is calculated according to the target difference detection information. The detection result of the target experiment index is determined based on the difference data.

12. The experimental data processing method of claim 11, wherein, The difference data of the target experiment index is calculated according to the target difference detection information, and the calculating comprises the following steps. A target difference regression coefficient and a target variance of the target difference regression coefficient are extracted from the target difference detection information. An experiment-after difference regression coefficient is identified from the target difference regression coefficient, and the experiment-after difference regression coefficient is fused with an experiment cumulative day number to obtain modified experiment-after difference information. Difference significance probability of the target experiment index is calculated based on the target difference regression coefficient and the target variance, and the modified experiment-after difference information and the difference significance probability are taken as the difference data.

13. The experimental data processing method of claim 12, wherein, The difference significance probability of the target experiment index is calculated based on the target difference regression coefficient and the target variance, and the calculating comprises the following steps. The target detection statistic of the target experiment index is obtained by fusing the target difference regression coefficient and the target variance. The basic probability information of the detection statistic not exceeding the target detection statistic is obtained, and the target probability information of the detection statistic exceeding the target detection statistic is calculated based on the basic probability information. The difference significance probability information of the target experiment index is determined based on the target probability information. The experiment-before difference significance probability and the modified experiment-after difference significance probability are extracted from the difference significance probability information as the difference significance probability of the target experiment index.

14. The experimental data processing method of claim 11, wherein, The detection result of the target experiment index is determined based on the difference data, and the determining comprises the following steps. The experiment-before difference significance probability is extracted from the difference data. When the experiment-before difference significance probability does not exceed a preset probability threshold, the difference data is taken as the detection result of the target experiment index. When the experiment-before difference significance probability exceeds the preset probability threshold, experiment-after experiment index data corresponding to the target experiment index is detected to obtain the detection result.

15. The experimental data processing method according to any one of claims 1 to 10, characterized in that, The experiment index data set is obtained, and the obtaining comprises the following steps. At least one current experiment index data of at least one experiment object at a current experiment day is obtained. Corresponding experiment index cumulative data is updated based on the current experiment index data. Daily average cumulative values of each updated experiment index cumulative data of the experiment object are calculated to obtain the experiment index data set.

16. An experimental data processing method, characterized by, The method comprises the following steps. A difference detection request for a target experiment index is sent to a server. A detection result of the target experiment index returned by the server is received, wherein the detection result is obtained by using the experiment data processing method in any one of claims 1 to 15. An experiment data processing page is displayed, and the experiment data processing page comprises a difference significance display control corresponding to a type of the detection result. In response to a display operation on the difference significance display control, the detection result of the target experimental index is displayed.

17. The experimental data processing method of claim 16, wherein, The experimental data processing page is displayed, including: When the type of the detection result is the pre-experiment difference, the corrected post-experiment difference significance probability is extracted from the detection result, and the experimental data processing page is displayed based on the corrected post-experiment difference significance probability; When the type of the detection result is the pre-experiment difference, the difference significance display control corresponding to the post-experiment difference is obtained, and the experimental data processing page is displayed based on the difference significance display control corresponding to the post-experiment difference.

18. The experimental data processing method of claim 17, wherein, The experimental data processing page is displayed based on the corrected post-experiment difference significance probability, including: When the corrected post-experiment difference significance probability does not exceed the preset significance probability threshold, it is determined that the correction type of the target experimental index is a correction significant type, and the difference significance display control corresponding to the correction significant type is obtained to display the experimental data processing page; When the corrected post-experiment difference significance probability exceeds the preset significance probability threshold, it is determined that the correction type of the target experimental index is a correction non-significant type, and the difference significance display control corresponding to the correction non-significant type is obtained to display the experimental data processing page.

19. An experimental data processing apparatus, characterized by, It includes: The acquisition unit is used to acquire an experimental index data set, and the experimental index data set includes experimental index data of at least one experimental index of at least one experimental object under a current experiment; The construction unit is used to construct a to-be-processed experimental data set for the experimental index according to the experimental grouping and data type of the experimental index data; The identification unit is used to identify a difference regression coefficient of each experimental index in the to-be-processed experimental data set based on a preset individual difference parameter; The calculation unit is used to filter panel experimental data in the to-be-processed experimental data set, and calculate the variance of the difference regression coefficient based on the panel experimental data to obtain difference detection information of each experimental index; The detection unit is used to perform difference detection on a target experimental index according to the difference detection information when a difference detection request for the target experimental index is received, and send a detection result to a terminal.

20. An experimental data processing apparatus, characterized by, It includes: The sending unit is used to send a difference detection request for a target experimental index to a server; The receiving unit is used to receive a detection result of the target experimental index returned by the server; wherein the detection result is obtained by the experimental data processing device of claim 19; The first display unit is used to display an experimental data processing page, and the experimental data processing page includes a difference significance display control corresponding to the type of the detection result; The second display unit is used to display the detection result of the target experimental index in response to a display operation on the difference significance display control.

21. An electronic device, comprising: It includes a processor and a memory, the memory stores an application program, and the processor is used to run the application program in the memory to execute the method of any one of claims 1-18. It includes a processor and a memory, the memory stores an application program, and the processor is used to run the application program in the memory to execute the method of any one of claims 1-18.

22. A computer-readable storage medium, characterized in that, The computer readable storage medium stores a plurality of instructions adapted to be loaded by the processor to perform the method according to any one of claims 1-18.

23. A computer program product, characterised in that, The computer program product comprises computer instructions stored in the computer readable storage medium; the processor of the computer device reads the computer instructions from the computer readable storage medium, and the processor executes the computer instructions, so that the computer device performs the method according to any one of claims 1-18.

Images