Data analysis method and device, storage medium and electronic equipment

Abstract

According to the data analysis method and device, the storage medium and the electronic equipment provided by the invention, the historical browsing data corresponding to the target recommendation information can be obtained, and the historical browsing data is input into the pre-trained prediction model, so that the prediction model predicts the browsing data corresponding to the target recommendation information at the specified time in the future; the method comprises the steps of taking target recommendation information as standard browsing data, obtaining actual browsing data of the target recommendation information at specified time, determining deviation between the actual browsing data and the standard browsing data, determining the actual browsing data as abnormal data when the deviation is greater than preset deviation, and performing data analysis on the abnormal data to obtain an analysis result.

Description

Technical Field

[0001] The present invention relates to the field of computer technology, and in particular to a data analysis method, device, storage medium and electronic device. Background Art

[0002] At present, in order to meet the personalized needs of users, enterprises can recommend information that meets the needs of users to improve the user experience. For example, the user's purchase behavior can be analyzed to recommend the user the product link that the user is interested in. For another example, the corresponding advertisement can be pushed to the user based on the historical browsing record information authorized by the user. After recommending the corresponding information to the user, the user's browsing data on the information can often be collected, and the collected browsing data can be analyzed to recommend the user the content that the user is more interested in, thereby further improving the user experience.

[0003] However, in the prior art, manual analysis is often used to analyze browsing data, resulting in low execution efficiency. Summary of the invention

[0004] This specification provides a data analysis method, device, storage medium and electronic device to partially solve the above-mentioned problems existing in the prior art.

[0005] This manual adopts the following technical solutions:

[0006] This specification provides a data analysis method, including:

[0007] Acquire historical browsing data corresponding to the target recommendation information, wherein the historical browsing data is used to reflect the historical browsing status of the target recommendation information;

[0008] Inputting the historical browsing data into a pre-trained prediction model so that the prediction model predicts browsing data corresponding to the target recommendation information at a specified time in the future as standard browsing data;

[0009] Acquire actual browsing data of the target recommendation information at the specified time, and determine a deviation between the actual browsing data and the standard browsing data;

[0010] When the deviation is greater than a preset deviation, the actual browsing data is determined to be abnormal data, and data analysis is performed on the abnormal data to obtain an analysis result.

[0011] Optionally, the step of obtaining historical browsing data corresponding to the target recommendation information specifically includes:

[0012] Obtain the initial browsing data of the target recommendation information in the historical period;

[0013] The initial browsing data is preprocessed to use the preprocessed data as the historical browsing data corresponding to the acquired target recommendation information, wherein the preprocessing includes at least one of deduplication processing, outlier processing, and standardization processing.

[0014] Optionally, the method further comprises:

[0015] When the deviation is not greater than the preset deviation, determining a reasonable browsing data interval according to the historical browsing data, and judging whether the actual browsing data is within the reasonable browsing data interval;

[0016] If not, the actual browsing data is regarded as abnormal data, and data analysis is performed on the abnormal data to obtain an analysis result.

[0017] Optionally, the step of performing data analysis on the abnormal data to obtain analysis results specifically includes:

[0018] Obtaining the recommended strategy information corresponding to the abnormal data in each business dimension;

[0019] For the recommendation strategy information corresponding to each business dimension, the recommendation strategy information corresponding to the business dimension is matched with the historical recommendation strategy information under the business dimension, and when it is determined that the recommendation strategy information corresponding to the business dimension does not match the historical recommendation strategy information under the business dimension, the recommendation strategy information corresponding to the business dimension is regarded as abnormal recommendation strategy information;

[0020] The abnormal recommendation strategy information is used as abnormal cause information of the abnormal data, and an analysis result is obtained according to the abnormal cause information.

[0021] Optionally, the method further comprises:

[0022] Inputting the abnormal recommendation strategy information and prompt statements into a preset macro model, so as to generate adjustment strategy information for avoiding the abnormal data through the macro model;

[0023] The abnormal recommendation strategy information is adjusted according to the adjustment strategy information so as to perform business according to the adjusted strategy information.

[0024] Optionally, the step of pre-training the prediction model specifically includes:

[0025] Obtaining sample browsing information and tag browsing data corresponding to the sample browsing information;

[0026] Inputting the sample browsing information into a prediction model to be trained, so that the prediction model to be trained determines the predicted browsing data corresponding to the sample browsing information;

[0027] A loss value is determined according to a deviation between the predicted browsing data and the label browsing data, and the prediction model to be trained is trained according to the loss value, wherein there is a positive correlation between the deviation and the loss value.

[0028] Optionally, the step of performing data analysis on the abnormal data to obtain analysis results specifically includes:

[0029] On the premise that the browsing situation reflected by the abnormal data is a normal browsing situation, predicting the group distribution of the user group that can generate the browsing situation reflected by the abnormal data after receiving the target recommendation information as a reference group distribution;

[0030] determining a group distribution of a user group that generates a browsing situation reflected by the actual browsing data after receiving the target recommendation information, as the actual group distribution;

[0031] According to the actual group distribution and the reference group distribution, abnormal recommendation strategy information is determined from each recommendation strategy information corresponding to each business dimension of the abnormal data, so as to obtain an analysis result according to the abnormal recommendation strategy information.

[0032] This specification provides a data analysis device, comprising:

[0033] A first acquisition module: used to acquire historical browsing data corresponding to target recommendation information, wherein the historical browsing data is used to reflect the historical browsing status of the target recommendation information;

[0034] Prediction module: used for inputting the historical browsing data into a pre-trained prediction model, so that the prediction model predicts the browsing data corresponding to the target recommendation information at a specified time in the future as standard browsing data;

[0035] A second acquisition module: used for acquiring actual browsing data of the target recommendation information at the specified time, and determining a deviation between the actual browsing data and the standard browsing data;

[0036] Analysis module: used for determining that the actual browsing data is abnormal data when the deviation is greater than a preset deviation, and performing data analysis on the abnormal data to obtain an analysis result.

[0037] This specification provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, the above-mentioned data analysis method is implemented.

[0038] This specification provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the above-mentioned data analysis method when executing the program.

[0039] At least one of the above technical solutions adopted in this specification can achieve the following beneficial effects:

[0040] The data analysis method provided in this specification can obtain historical browsing data corresponding to target recommendation information, input the historical browsing data into a pre-trained prediction model, so that the prediction model predicts the browsing data corresponding to the target recommendation information at a specified time in the future as standard browsing data, obtain the actual browsing data of the target recommendation information at the specified time, and determine the deviation between the actual browsing data and the standard browsing data. When the deviation is greater than the preset deviation, the actual browsing data is determined to be abnormal data, and data analysis is performed on the abnormal data to obtain analysis results.

[0041] It can be seen from this that the above method can judge and analyze whether the actual browsing data is abnormal data based on the deviation between the standard browsing data predicted by the prediction model and the actual browsing data, that is, there is no need to use manual analysis to analyze the browsing data like the existing technology, thereby greatly improving the execution efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] The drawings described herein are used to provide a further understanding of this specification and constitute a part of this specification. The illustrative embodiments and descriptions of this specification are used to explain this specification and do not constitute an improper limitation on this specification. In the drawings:

[0043] Figure 1 A schematic diagram of a data analysis method provided in this specification;

[0044] Figure 2 A schematic diagram of a data analysis process provided in this specification;

[0045] Figure 3 A schematic diagram of a data analysis process provided in this specification;

[0046] Figure 4 A schematic diagram of a data analysis device provided in this specification;

[0047] Figure 5 A method corresponding to the Figure 1 Schematic block diagram of an electronic device. DETAILED DESCRIPTION

[0048] In order to make the purpose, technical solutions and advantages of this specification more clear, the technical solutions of this specification will be clearly and completely described below in combination with the specific embodiments of this specification and the corresponding drawings. Obviously, the described embodiments are only part of the embodiments of this specification, not all of them. Based on the embodiments in this specification, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this specification.

[0049] At present, in order to meet the personalized needs of users, enterprises can recommend information that meets the needs of users to improve the user experience. For example, the user's purchasing behavior can be analyzed to recommend the user the product link that the user is interested in. For another example, the corresponding advertisement can be pushed to the user based on the historical browsing record information authorized by the user. After recommending the corresponding information to the user, the user's browsing data on the information can often be collected, and the collected browsing data can be analyzed to recommend the user with more interesting content, thereby further improving the user experience. However, in the prior art, the browsing data is often analyzed by manual analysis, resulting in low execution efficiency.

[0050] Therefore, this specification provides a data analysis method, which can effectively solve the problems in the prior art.

[0051] The technical solutions provided by the embodiments of this specification are described in detail below in conjunction with the accompanying drawings.

[0052] Figure 1 A flow chart of a data analysis method provided in this specification includes the following steps:

[0053] S101: Acquire historical browsing data corresponding to target recommendation information, where the historical browsing data is used to reflect the historical browsing status of the target recommendation information.

[0054] The execution subject of the data analysis method involved in this specification can be a terminal device such as a desktop computer, a laptop computer, or a client installed in the terminal device, or a server. The following only takes the server as an example to illustrate the data analysis method in the embodiment of this specification.

[0055] In this specification, the server can obtain the historical browsing data corresponding to the target recommendation information, and then determine the abnormal data therefrom, and analyze the abnormal data to obtain the corresponding analysis results. Among them, the target recommendation information can be the link information of the product that the user is interested in, or the target recommendation information can also be the corresponding advertising information pushed to the user based on the historical browsing record information authorized by the user. Correspondingly, the browsing data can be the click rate of the user corresponding to the target recommendation information, the conversion rate of the user corresponding to the target recommendation information, and other data that can reflect the browsing situation of the target recommendation information.

[0056] First, the server can obtain the historical browsing data corresponding to the target recommendation information, so as to perform data analysis on the actual browsing data according to the obtained historical browsing data to obtain the analysis result. The historical browsing data can reflect the historical browsing situation of the target recommendation information.

[0057] Of course, in the process of obtaining the historical browsing data corresponding to the target recommendation information, the data can also be preprocessed. Specifically, the server can first obtain the initial browsing data of the target recommendation information in the historical period, and then preprocess the initial browsing data, and use the preprocessed data as the historical browsing data corresponding to the obtained target recommendation information. For example: the initial browsing data can be deduplicated. For another example, the initial browsing data can be processed for outliers. For another example, the initial browsing data can be standardized (such as: converting the initial browsing data into data of the same format).

[0058] S102: Inputting the historical browsing data into a pre-trained prediction model, so that the prediction model predicts browsing data corresponding to the target recommendation information at a specified time in the future as standard browsing data.

[0059] In this specification, the browsing data corresponding to the target recommendation information at a specified time in the future can be predicted by a pre-trained prediction model, so as to determine whether the actual browsing data at the specified time is abnormal data based on the predicted data.

[0060] Specifically, in the process of training the prediction model to be trained, the sample browsing information and the label browsing data corresponding to the sample browsing information can be obtained first. Among them, the sample browsing information can be understood as the sample browsing data from the sample data set. The server can input the sample browsing information into the prediction model to be trained, and the model can determine the predicted browsing data corresponding to the sample browsing information. Furthermore, the loss value can be determined based on the deviation between the predicted browsing data and the label browsing data, wherein the deviation and the loss value are positively correlated. The server can train the prediction model with gradually reducing the loss value as the training goal. Of course, in the actual application process, different training methods can be used to train the prediction model according to actual needs, such as: reinforcement learning, semi-supervised learning, etc., which will not be elaborated in this manual.

[0061] After the prediction model is trained, the server can input the historical browsing data corresponding to the target recommendation information into the prediction model, so that the prediction model can predict the browsing data corresponding to the target recommendation information at a specified time in the future and use it as the standard browsing data. For example, the prediction model can predict the browsing data corresponding to the target recommendation information in the next week, or predict the browsing data corresponding to the target recommendation information on a certain day in the future.

[0062] S103: Acquire actual browsing data of the target recommendation information at the specified time, and determine a deviation between the actual browsing data and the standard browsing data.

[0063] S104: When the deviation is greater than a preset deviation, the actual browsing data is determined to be abnormal data, and data analysis is performed on the abnormal data to obtain an analysis result.

[0064] After predicting the browsing data at a specified time in the future, the server can obtain the actual browsing data of the target recommendation information at the specified time and determine the deviation between the actual browsing data and the standard browsing data. When the determined deviation is less than the preset deviation, the actual browsing data can be used as abnormal data.

[0065] Of course, in addition to judging abnormal data according to standard browsing data, it is also possible to judge whether the actual browsing data is abnormal data according to the reasonable range of browsing data. Specifically, when the deviation between the actual browsing data and the standard browsing data is not greater than the preset deviation, the server can further judge whether the actual browsing data is within the reasonable range of browsing data. If not, the actual browsing data can be regarded as abnormal data. Correspondingly, if the actual browsing data is within the reasonable range of browsing data, the actual browsing data can be regarded as normal data.

[0066] The above-mentioned reasonable browsing data interval can be preset by the administrator. Of course, the reasonable browsing data interval can also be determined by the server according to the historical browsing data obtained. Specifically, since the historical browsing data obtained in actual applications are often multiple, the minimum historical browsing data among the historical browsing data can be used as the left end point of the reasonable browsing interval, and the maximum historical browsing data among the historical browsing data can be used as the right end point of the reasonable browsing interval. For example, the reasonable browsing data interval can be a conversion rate of 45%---a conversion rate of 65%.

[0067] In addition, in this specification, it is also possible to directly determine whether the actual browsing data is abnormal data based on the reasonable range of browsing data. Specifically, the server can directly determine whether the actual browsing data of the target recommended information is within the reasonable range of browsing data. If not, the actual browsing data can be regarded as abnormal data. Of course, when the actual browsing data of the target recommended information is within the reasonable range of browsing data, further, the deviation between the actual browsing data and the corresponding standard browsing data predicted by the prediction model can be determined. If the deviation is greater than the preset deviation, the actual browsing data is determined to be abnormal data. On the contrary, if the deviation is not greater than the preset deviation, the actual browsing data is determined to be normal data.

[0068] When the actual browsing data is determined to be abnormal data, the abnormal data can be analyzed to obtain analysis results. For example, the click rate of users corresponding to the target recommendation information in a specified time period can be displayed in the form of a line graph, and the abnormal click rate (such as: abnormally high click rate) in the line graph can be bolded.

[0069] In addition, the reasons for abnormal data can also be analyzed. Specifically, the server can obtain the recommendation strategy information corresponding to the abnormal data in each business dimension, and the recommendation strategy information corresponding to each business dimension can include the specific time of the recommendation information, the cost of the recommendation information (such as: content design cost, information delivery cost, etc.), the recommended user group, etc.

[0070] Furthermore, for the recommendation strategy information corresponding to each business dimension, the recommendation strategy information corresponding to the business dimension can be matched with the historical recommendation strategy information under the business dimension, and when it is determined that the recommendation strategy information corresponding to the business dimension does not match the historical recommendation strategy information under the business dimension, the recommendation strategy information corresponding to the business dimension is used as abnormal recommendation strategy information. Furthermore, the abnormal recommendation strategy information can be used as abnormal cause information of abnormal data, and the analysis result can be obtained according to the abnormal cause information.

[0071] For example, when the user click-through rate in the browsing data corresponding to the target recommended information on a certain day is abnormally low, the server can match the specific time of the recommended information on that day with the average time of the recommended information in the past month. If the specific time of the recommended information on that day is significantly later than the average time, it can be determined that the abnormal recommendation strategy information is the time of the recommended information on that day, that is, the abnormally low click-through rate is caused by the late recommendation time.

[0072] For another example, when the user click rate in the browsing data corresponding to the target recommendation information on a certain day is abnormally high, the server can compare the cost spent on recommending the target recommendation information on that day with the average cost spent on recommending the target recommendation information in the past month. If the cost spent on recommending the target recommendation information on that day is significantly higher than the average cost, it can be determined that the abnormal recommendation strategy information is the cost spent on recommending the target recommendation information on that day. Furthermore, the server can also compare the delivery information cost in the cost spent on recommending the target recommendation information on that day with the average delivery information cost spent on recommending the target recommendation information in the past month. If the delivery information cost is significantly higher than the average delivery information cost, it can be further determined that the abnormal recommendation strategy information is the delivery information cost spent on recommending the target recommendation information on that day, that is, the abnormally high click rate on that day was caused by the abnormally high delivery information cost.

[0073] In addition, the server can also generate adjustment strategy information to avoid abnormal data. Specifically, the abnormal recommendation strategy information and the corresponding prompt statement can be input into a preset large model (such as: GPT-4), and then the adjustment strategy information to avoid abnormal data can be generated through the large model. For example, the time of the abnormal recommendation information and the prompt statement "Please generate a corresponding adjustment strategy based on the current abnormal recommendation strategy information" can be input, and then the large model can generate the adjustment strategy information "Please appropriately advance the time of recommending the corresponding information to avoid abnormal data from appearing again". Of course, in addition to generating the corresponding adjustment strategy information through the above-mentioned large model, the abnormal recommendation strategy information can also be analyzed through a pre-set program code to generate the corresponding adjustment strategy information.

[0074] Furthermore, the server can adjust the abnormal recommendation strategy information according to the adjustment strategy information, so as to perform the business according to the adjusted strategy information. For example, when the adjustment strategy information is "please appropriately advance the time of the recommendation information to avoid abnormal data from appearing again", the abnormal recommendation time can be advanced according to the adjustment strategy information, so as to avoid the situation where the user's click rate is abnormally low. Correspondingly, the target recommendation information recommended according to the adjusted recommendation strategy will also be more in line with the user's own needs.

[0075] Of course, in actual application, abnormal data may also appear due to abnormal distribution of the recommended user group. Therefore, in the process of analyzing the abnormal data, data analysis can also be performed on the user group corresponding to the target recommendation information.

[0076] Specifically, when it is determined that the actual browsing data is abnormal data, the browsing situation reflected by the abnormal data can be first regarded as the normal browsing situation, so that on the premise that it is the normal browsing situation, the group distribution of the user group that can generate the browsing situation reflected by the abnormal data after receiving the target recommendation information is predicted as the reference group distribution. For example, the actual browsing data can be input into a preset parsing model so that the parsing model predicts the reference group distribution that can generate the browsing situation reflected by the actual browsing data. For another example, the reference group distribution that can generate the browsing situation reflected by the actual browsing data can be parsed according to the browsing situation reflected by the abnormal data through a preset program code.

[0077] Then, the group distribution of the user group whose browsing situation is reflected by the actual browsing data generated after receiving the target recommendation information can be determined and used as the actual group distribution. The server can determine the abnormal recommendation strategy information from the corresponding recommendation strategy information under each business dimension of the abnormal data according to the actual group distribution and the reference group distribution, so as to obtain the analysis result according to the abnormal recommendation strategy information.

[0078] For example, when an abnormal click rate of 10% occurs, it can be predicted that the reference group distribution corresponding to a normal click rate of 10% is: 40% of males aged 18-28, 35% of females aged 18-28, 15% of males aged 29-40, and 10% of females aged 29-40. The actual group distribution corresponding to the abnormal click rate of 10% is: 30% of males aged 18-28, 45% of females aged 18-28, 10% of males aged 29-40, and 15% of males aged 29-40. Then, by comparing the actual group distribution with the reference group distribution, it can be determined that there is an abnormality in the proportion of males and females in the two age groups in the actual group distribution. For this reason, the corresponding user group distribution can be adjusted. At the same time, the content of the target recommendation information can also be adjusted and optimized, so as to provide the corresponding user groups with information that better meets their needs.

[0079] Of course, in the actual application process, the target recommendation information can be placed on different delivery platforms according to actual needs. Correspondingly, the corresponding browsing data can be obtained from different delivery platforms to perform data analysis, and the analysis results can also be returned to the corresponding managers in the form of word documents, pdf documents, etc. for their review.

[0080] In order to further explain the above data analysis method, the overall process of data analysis will be explained in the form of a legend. For details, please refer to the following Figure 2 , Figure 3 .

[0081] Figure 2 A schematic diagram of a data analysis process provided in this manual.

[0082] Depend on Figure 2 It can be seen that in the process of data analysis, the historical browsing data corresponding to the target recommendation information can be input into the prediction model so that the prediction model predicts the corresponding standard browsing data. Then, the deviation between the corresponding actual browsing data and the standard browsing data can be determined, and it can be judged whether the deviation is less than the preset deviation. When the deviation is not less than the preset deviation, the actual browsing data can be regarded as abnormal data, and data analysis is performed on the abnormal data to obtain the analysis result. On the contrary, when the deviation is less than the preset deviation, it can be further judged whether the actual browsing data is abnormal again. When the actual browsing data is within the reasonable range of browsing data, the actual browsing data is regarded as normal data; when the actual browsing data is not within the reasonable range of browsing data, the actual browsing data is regarded as abnormal data, and data analysis is performed on the abnormal data to obtain the analysis result.

[0083] Figure 3 A schematic diagram of a data analysis process provided in this manual.

[0084] Depend on Figure 3It can be seen that for the actual browsing data, it is also possible to determine whether the actual browsing data is abnormal data by only performing an abnormal judgment once, so as to perform subsequent data analysis. Specifically, when the actual browsing data is within a reasonable range of browsing data, the actual browsing data can be regarded as normal data; when the actual browsing data is not within a reasonable range of browsing data, the actual browsing data can be regarded as abnormal data, and data analysis can be performed on the abnormal data to obtain analysis results. Alternatively, the corresponding standard browsing data can be predicted based on the historical browsing data through a prediction model. When the deviation between the actual browsing data and the standard browsing data is less than a preset deviation, the actual browsing data can be regarded as normal data; when the deviation between the actual browsing data and the standard browsing data is not less than a preset deviation, the actual browsing data can be regarded as abnormal data, and data analysis can be performed on the abnormal data to obtain analysis results.

[0085] It can be seen from this that the above method can judge and analyze whether the actual browsing data is abnormal data based on the deviation between the standard browsing data predicted by the prediction model and the actual browsing data, that is, there is no need to use manual analysis to analyze the browsing data like the existing technology, thereby greatly improving the execution efficiency.

[0086] In addition, in the above method, the abnormal cause information of the abnormal data can also be analyzed, so that the abnormal recommendation strategy information can be adjusted accordingly according to the analyzed abnormal cause information, and then the information that better meets the user's own needs can be recommended to the user according to the adjusted recommendation strategy, thereby improving the execution effect.

[0087] The above are one or more implementation methods of this specification. Based on the same idea, this specification also provides a corresponding data analysis device, such as Figure 4 shown.

[0088] Figure 4 A schematic diagram of a data analysis device provided in this specification includes:

[0089] The first acquisition module 401 is used to acquire historical browsing data corresponding to the target recommendation information, where the historical browsing data is used to reflect the historical browsing status of the target recommendation information;

[0090] Prediction module 402: used to input the historical browsing data into a pre-trained prediction model, so that the prediction model predicts browsing data corresponding to the target recommendation information at a specified time in the future as standard browsing data;

[0091] The second acquisition module 403 is used to acquire the actual browsing data of the target recommendation information at the specified time, and determine the deviation between the actual browsing data and the standard browsing data;

[0092] Analysis module 404: for determining that the actual browsing data is abnormal data when the deviation is greater than a preset deviation, and performing data analysis on the abnormal data to obtain an analysis result.

[0093] Optionally, the first acquisition module 401 is specifically used to: obtain initial browsing data of target recommendation information in a historical period; preprocess the initial browsing data to use the preprocessed data as historical browsing data corresponding to the acquired target recommendation information, wherein the preprocessing includes at least one of deduplication processing, outlier processing, and standardization processing.

[0094] Optionally, the device further includes: a judgment module 405;

[0095] The judgment module 405 is specifically used for: when the deviation is not greater than the preset deviation, determining a reasonable browsing data interval according to the historical browsing data, and judging whether the actual browsing data is within the reasonable browsing data interval; if not, treating the actual browsing data as abnormal data, and performing data analysis on the abnormal data to obtain an analysis result.

[0096] Optionally, the analysis module 404 is specifically used to: obtain the recommendation strategy information corresponding to the abnormal data under each business dimension; for the recommendation strategy information corresponding to each business dimension, match the recommendation strategy information corresponding to the business dimension with the historical recommendation strategy information under the business dimension, and when it is determined that the recommendation strategy information corresponding to the business dimension does not match the historical recommendation strategy information under the business dimension, use the recommendation strategy information corresponding to the business dimension as abnormal recommendation strategy information; use the abnormal recommendation strategy information as abnormal cause information for the occurrence of the abnormal data, and obtain analysis results based on the abnormal cause information.

[0097] Optionally, the method further includes: an adjustment module 406;

[0098] The adjustment module 406 is specifically used to: input the abnormal recommendation strategy information and prompt statement into a preset big model, so as to generate adjustment strategy information for avoiding the occurrence of the abnormal data through the big model; adjust the abnormal recommendation strategy information according to the adjustment strategy information, so as to perform business according to the adjusted strategy information.

[0099] Optionally, the prediction module 402 is specifically used to: obtain sample browsing information and label browsing data corresponding to the sample browsing information; input the sample browsing information into the prediction model to be trained so that the prediction model to be trained determines the predicted browsing data corresponding to the sample browsing information; determine a loss value based on the deviation between the predicted browsing data and the label browsing data, and train the prediction model to be trained based on the loss value, wherein the deviation and the loss value are positively correlated.

[0100] Optionally, the analysis module 404 is specifically used to: on the premise that the browsing situation reflected by the abnormal data is a normal browsing situation, predict the group distribution of the user group that can generate the browsing situation reflected by the abnormal data after receiving the target recommendation information, as a reference group distribution; determine the group distribution of the user group that generates the browsing situation reflected by the actual browsing data after receiving the target recommendation information, as the actual group distribution; according to the actual group distribution and the reference group distribution, determine the abnormal recommendation strategy information from the corresponding recommendation strategy information under each business dimension of the abnormal data, so as to obtain the analysis result according to the abnormal recommendation strategy information.

[0101] This specification also provides a computer-readable storage medium, which stores a computer program, which can be used to execute the above Figure 1 The data analysis methods shown.

[0102] This manual also provides Figure 5 The one shown corresponds to Figure 1 A schematic diagram of the electronic device. Figure 5 As shown, at the hardware level, the electronic device includes a processor, an internal bus, a network interface, a memory, and a non-volatile memory, and may also include other hardware required for the business. The processor reads the corresponding computer program from the non-volatile memory into the memory and then runs it to achieve the above Figure 1 The data analysis methods shown.

[0103] Of course, in addition to software implementation, this specification does not exclude other implementation methods, such as logic devices or a combination of software and hardware, etc., that is to say, the executor of the following processing flow is not limited to each logic unit, but can also be hardware or logic devices.

[0104] In the 1990s, improvements to a technology could be clearly distinguished as hardware improvements (for example, improvements to the circuit structure of diodes, transistors, switches, etc.) or software improvements (improvements to the method flow). However, with the development of technology, many improvements to the method flow today can be regarded as direct improvements to the hardware circuit structure. Designers almost always obtain the corresponding hardware circuit structure by programming the improved method flow into the hardware circuit. Therefore, it cannot be said that an improvement in a method flow cannot be implemented using a hardware entity module. For example, a programmable logic device (PLD) (such as a field programmable gate array (FPGA)) is such an integrated circuit whose logical function is determined by the user's programming of the device. Designers can "integrate" a digital system on a PLD by programming it themselves, without having to ask a chip manufacturer to design and produce a dedicated integrated circuit chip. Moreover, nowadays, instead of manually making integrated circuit chips, this kind of programming is mostly implemented by "logic compiler" software, which is similar to the software compiler used when developing and writing programs, and the original code before compilation must also be written in a specific programming language, which is called hardware description language (HDL). There is not only one HDL, but many kinds, such as ABEL (Advanced Boolean Expression Language), AHDL (Altera Hardware Description Language), Confluence, CUPL (Cornell University Programming Language), HDCal, JHDL (Java Hardware Description Language), Lava, Lola, MyHDL, PALASM, RHDL (Ruby Hardware Description Language), etc. The most commonly used ones are VHDL (Very-High-Speed ​​Integrated Circuit Hardware Description Language) and Verilog. Those skilled in the art should also know that it is only necessary to program the method flow slightly in the above-mentioned hardware description languages ​​and program it into the integrated circuit, and then it is easy to obtain the hardware circuit that implements the logic method flow.

[0105] The controller can be implemented in any appropriate manner, for example, the controller can take the form of a microprocessor or processor and a computer-readable medium storing a computer-readable program code (such as software or firmware) that can be executed by the (micro)processor, a logic gate, a switch, an application-specific integrated circuit (ASIC), a programmable logic controller, and an embedded microcontroller. Examples of controllers include, but are not limited to, the following microcontrollers: ARC 625D, Atmel AT91SAM, Microchip PIC18F26K20, and Silicone Labs C8051F320. The memory controller can also be implemented as part of the control logic of the memory. Those skilled in the art also know that in addition to implementing the controller in a purely computer-readable program code manner, the controller can be implemented in the form of a logic gate, a switch, an application-specific integrated circuit, a programmable logic controller, and an embedded microcontroller by logically programming the method steps. Therefore, this controller can be considered as a hardware component, and the devices included therein for implementing various functions can also be regarded as structures within the hardware component. Or even, the devices for implementing various functions can be regarded as both software modules for implementing the method and structures within the hardware component.

[0106] The systems, devices, modules or units described in the above embodiments may be implemented by computer chips or entities, or by products with certain functions. A typical implementation device is a computer. Specifically, the computer may be, for example, a personal computer, a laptop computer, a cellular phone, a camera phone, a smart phone, a personal digital assistant, a media player, a navigation device, an email device, a game console, a tablet computer, a wearable device, or a combination of any of these devices.

[0107] For the convenience of description, the above device is described in various units according to their functions. Of course, when implementing this specification, the functions of each unit can be implemented in the same or multiple software and / or hardware.

[0108] Those skilled in the art will appreciate that the embodiments of this specification may be provided as methods, systems, or computer program products. Therefore, this specification may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, this specification may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program code.

[0109] This specification is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of this specification. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0110] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0111] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0112] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0113] The memory may include non-permanent storage in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. The memory is an example of a computer-readable medium.

[0114] Computer readable media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer readable media does not include temporary computer readable media (transitory media), such as modulated data signals and carrier waves.

[0115] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

[0116] Those skilled in the art will appreciate that the embodiments of this specification may be provided as methods, systems or computer program products. Therefore, this specification may take the form of a complete hardware embodiment, a complete software embodiment or an embodiment combining software and hardware. Moreover, this specification may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0117] This specification may be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types. This specification may also be practiced in distributed computing environments where tasks are performed by remote processing devices connected through a communication network. In a distributed computing environment, program modules may be located in local and remote computer storage media, including storage devices.

[0118] Each embodiment in this specification is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the system embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.

[0119] The above description is only an embodiment of the present specification and is not intended to limit the present specification. For those skilled in the art, the present specification may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present specification shall be included in the scope of the claims of the present specification.

Claims

1. A data analysis method, characterized in that: include: Acquire historical browsing data corresponding to the target recommendation information, wherein the historical browsing data is used to reflect the historical browsing status of the target recommendation information; Inputting the historical browsing data into a pre-trained prediction model so that the prediction model predicts browsing data corresponding to the target recommendation information at a specified time in the future as standard browsing data; Acquire actual browsing data of the target recommendation information at the specified time, and determine a deviation between the actual browsing data and the standard browsing data; When the deviation is greater than a preset deviation, the actual browsing data is determined to be abnormal data, and data analysis is performed on the abnormal data to obtain an analysis result.

2. The method according to claim 1, characterized in that The step of obtaining the historical browsing data corresponding to the target recommendation information specifically includes: Obtain the initial browsing data of the target recommendation information in the historical period; The initial browsing data is preprocessed to use the preprocessed data as the historical browsing data corresponding to the acquired target recommendation information, wherein the preprocessing includes at least one of deduplication processing, outlier processing, and standardization processing.

3. The method according to claim 1, characterized in that The method further comprises: When the deviation is not greater than the preset deviation, determining a reasonable browsing data interval according to the historical browsing data, and judging whether the actual browsing data is within the reasonable browsing data interval; If not, the actual browsing data is regarded as abnormal data, and data analysis is performed on the abnormal data to obtain an analysis result.

4. The method according to claim 1 or 3, characterized in that The step of performing data analysis on the abnormal data to obtain analysis results specifically includes: Obtaining the recommended strategy information corresponding to the abnormal data in each business dimension; For the recommendation strategy information corresponding to each business dimension, the recommendation strategy information corresponding to the business dimension is matched with the historical recommendation strategy information under the business dimension, and when it is determined that the recommendation strategy information corresponding to the business dimension does not match the historical recommendation strategy information under the business dimension, the recommendation strategy information corresponding to the business dimension is regarded as abnormal recommendation strategy information; The abnormal recommendation strategy information is used as abnormal cause information of the abnormal data, and an analysis result is obtained according to the abnormal cause information.

5. The method according to claim 4, characterized in that The method further comprises: Inputting the abnormal recommendation strategy information and prompt statements into a preset macro model, so as to generate adjustment strategy information for avoiding the abnormal data through the macro model; The abnormal recommendation strategy information is adjusted according to the adjustment strategy information so as to perform business according to the adjusted strategy information.

6. The method according to claim 1, characterized in that The step of pre-training the prediction model specifically includes: Obtaining sample browsing information and tag browsing data corresponding to the sample browsing information; Inputting the sample browsing information into a prediction model to be trained, so that the prediction model to be trained determines the predicted browsing data corresponding to the sample browsing information; A loss value is determined according to a deviation between the predicted browsing data and the label browsing data, and the prediction model to be trained is trained according to the loss value, wherein there is a positive correlation between the deviation and the loss value.

7. The method according to claim 1, characterized in that The step of performing data analysis on the abnormal data to obtain analysis results specifically includes: On the premise that the browsing situation reflected by the abnormal data is a normal browsing situation, predicting the group distribution of the user group that can generate the browsing situation reflected by the abnormal data after receiving the target recommendation information as a reference group distribution; determining a group distribution of a user group that generates a browsing situation reflected by the actual browsing data after receiving the target recommendation information, as the actual group distribution; According to the actual group distribution and the reference group distribution, abnormal recommendation strategy information is determined from each recommendation strategy information corresponding to each business dimension of the abnormal data, so as to obtain an analysis result according to the abnormal recommendation strategy information.

8. A data analysis device, characterized in that: include: A first acquisition module: used to acquire historical browsing data corresponding to target recommendation information, wherein the historical browsing data is used to reflect the historical browsing status of the target recommendation information; Prediction module: used for inputting the historical browsing data into a pre-trained prediction model, so that the prediction model predicts the browsing data corresponding to the target recommendation information at a specified time in the future as standard browsing data; A second acquisition module: used for acquiring actual browsing data of the target recommendation information at the specified time, and determining a deviation between the actual browsing data and the standard browsing data; Analysis module: used for determining that the actual browsing data is abnormal data when the deviation is greater than a preset deviation, and performing data analysis on the abnormal data to obtain an analysis result.

9. A computer-readable storage medium, characterized in that: The storage medium stores a computer program, and when the computer program is executed by a processor, the method according to any one of claims 1 to 7 is implemented.

10. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, the method according to any one of claims 1 to 7 is implemented.

Images