Data dimensionality reduction processing method, device, computer equipment and storage medium

Abstract

The present application relates to the field of big data processing, and in particular to a data dimension reduction processing method, device, computer equipment and storage medium. The method comprises: extracting relevant historical analysis models from a database, and parsing the historical analysis models to obtain initial features whose priority is greater than the target priority; obtaining the frequency of the initial features, and selecting the initial features whose frequency meets the requirements as target features; extracting initial data corresponding to the target features in multidimensional data; performing data processing of the same order of magnitude on different types of initial data to obtain standard data; performing nonlinear dimension reduction processing on the standard data to obtain target data of preset dimensions. The use of this method can improve the efficiency of subsequent data processing.

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence technology, and in particular to a data dimensionality reduction processing method, device, computer equipment and storage medium. Background Art

[0002] With the development of society, various types of data are increasing. In order to obtain more valuable data, more and more professionals have begun to conduct relevant research and analysis on various data.

[0003] However, for data with huge volume and numerous data types, directly analyzing and processing the data will consume a lot of processing time. This problem often leads to low processing efficiency of the hardware of the data processing system, which is time-consuming and resource-consuming. Summary of the invention

[0004] Based on this, it is necessary to provide a data dimensionality reduction processing method, device, computer equipment and storage medium that can improve the efficiency of subsequent data processing in response to the above technical problems.

[0005] A data dimension reduction processing method, the method comprising:

[0006] Extracting relevant historical analysis models from a database, and parsing the historical analysis models to obtain initial features with a priority greater than a target priority;

[0007] Obtaining the frequency of the initial feature, and selecting the initial feature whose frequency meets the requirement as the target feature;

[0008] Extracting initial data corresponding to the target feature from the multidimensional data;

[0009] Performing data processing of the same data level on different types of initial data to obtain standard data;

[0010] The standard data is subjected to nonlinear dimensionality reduction processing to obtain target data of a preset dimension.

[0011] In one embodiment, there are multiple historical analysis models, and the initial features whose priority is greater than the target priority are obtained by parsing the historical analysis models, including:

[0012] Analyzing each of the historical analysis models respectively to obtain initial features corresponding to each of the historical analysis models whose priorities are greater than the target priority;

[0013] The obtaining of the frequency of the initial feature and selecting the initial feature whose frequency meets the requirement as the target feature includes:

[0014] Determining the feature priority and frequency of occurrence of each of the initial features in each of the historical analysis models;

[0015] Comprehensively calculating the feature priority and the frequency of occurrence of each of the initial features in each of the historical analysis models to obtain the relative priority level of each of the initial features in each of the historical analysis models;

[0016] The relative priority levels of the initial features in the historical analysis models are compared, and the initial features whose relative priority levels meet the requirements are determined as the target features.

[0017] In one embodiment, after extracting the initial data corresponding to the target feature from the multidimensional data, the method further includes:

[0018] Acquire the data source of the initial data and the user address corresponding to the initial data, and divide the historical data in the initial data into local data and remote data based on the data source and the user address;

[0019] The performing data processing of the same data level on the different types of initial data to obtain standard data includes:

[0020] Performing data processing of the same data level on the different types of local data to obtain local standard data;

[0021] Performing data processing of the same data level on different types of remote data to obtain remote standard data;

[0022] The performing nonlinear dimensionality reduction processing on the standard data to obtain target data of preset dimensions includes:

[0023] Nonlinear dimensionality reduction processing is performed on the local standard data and the remote standard data respectively to obtain local target data and remote target data of preset dimensions.

[0024] In one embodiment, before performing data processing of the same data level on the different types of initial data to obtain standard data, the method further includes:

[0025] Dividing the initial data into cost data and frequency data according to data type;

[0026] The performing data processing of the same data level on the different types of initial data to obtain standard data includes:

[0027] Respectively obtain preset formulas corresponding to the cost data and the frequency data;

[0028] According to the preset formula corresponding to the expense data and the corresponding standard order of magnitude, the expense data of different data orders are processed at the same data order to obtain standard data corresponding to the expense data;

[0029] According to the preset formula corresponding to the frequency data and the corresponding standard order of magnitude, data processing of the same data order of magnitude is performed on the frequency data of different data orders to obtain standard data corresponding to the frequency data.

[0030] In one embodiment, after extracting the initial data corresponding to the target feature from the multidimensional data, the method further includes:

[0031] Performing data missing detection on data of remaining data types in the initial data according to data of at least one data type in the initial data;

[0032] When it is detected that the initial data is missing data, data of the data type is filled with data of the same data type as the missing data.

[0033] A data dimension reduction processing device, the device comprising:

[0034] A parsing module, used for extracting relevant historical analysis models from a database, and parsing the historical analysis models to obtain initial features having a priority greater than a target priority;

[0035] A target feature generation module is used to obtain the frequency of the initial feature and select the initial feature whose frequency meets the requirements as the target feature;

[0036] A data extraction module, used to extract initial data corresponding to the target feature from the multidimensional data;

[0037] A data processing module of the same data magnitude is used to perform data processing of the same data magnitude on different types of initial data to obtain standard data;

[0038] The dimensionality reduction processing module is used to perform nonlinear dimensionality reduction processing on the standard data to obtain target data of preset dimensions.

[0039] In one embodiment, there are multiple historical analysis models, and the parsing module is used to parse each of the historical analysis models respectively to obtain initial features corresponding to each of the historical analysis models whose priority is greater than the target priority;

[0040] The target feature generation module comprises:

[0041] A first determination submodule is used to determine the feature priority and the frequency of occurrence of the initial feature in the historical analysis model;

[0042] A calculation submodule, used for comprehensively calculating the feature priority and the frequency of occurrence of each of the initial features in each of the historical analysis models, to obtain the relative priority level of each of the initial features in each of the historical analysis models;

[0043] The comparison and determination submodule is used to compare the relative priority levels of the initial features in the historical analysis models, and determine the initial features whose relative priority levels meet the requirements as the target features.

[0044] In one embodiment, the device further comprises:

[0045] A data classification module, for obtaining the data source of the initial data and the user address corresponding to the initial data after the data extraction module extracts the initial data corresponding to the target feature in the multidimensional data, and classifying the historical data in the initial data into local data and remote data based on the data source and the user address;

[0046] The same data level data processing module comprises:

[0047] A first same data level data processing submodule, used for performing data processing of the same data level on different types of local data to obtain local standard data;

[0048] A second same data level data processing submodule is used to perform data processing of the same data level on different types of remote data to obtain remote standard data;

[0049] The dimension reduction processing module is used to perform nonlinear dimension reduction processing on the local standard data and the remote standard data respectively to obtain local target data and remote target data of preset dimensions.

[0050] A computer device comprises a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of any one of the above methods when executing the computer program.

[0051] A computer-readable storage medium stores a computer program, wherein the computer program implements the steps of any one of the above methods when executed by a processor.

[0052] The above-mentioned data dimensionality reduction processing method, device, computer equipment and storage medium obtain the target features through the historical analysis model, then extract the initial data corresponding to the target features in the multidimensional data, and obtain standard data after processing the data of the same data magnitude, and perform nonlinear dimensionality reduction processing on the standard data to obtain the target data of the preset dimension. The generated target data is generated based on the multidimensional data and is associated with the multidimensional data, so that the characteristics of the multidimensional data can be maintained, and then subsequent data processing and analysis can be performed through the target data. Compared with data processing and analysis using multidimensional data, it can save resource consumption when the system performs data processing and analysis, and can improve data processing efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Figure 1 This is an application scenario diagram of a data dimension reduction processing method in an embodiment;

[0054] Figure 2 A schematic diagram of a process flow of a data dimension reduction processing method in one embodiment;

[0055] Figure 3 A schematic flow chart of a target feature determination step in one embodiment;

[0056] Figure 4 Schematic diagram of a process of data missing detection step in one embodiment;

[0057] Figure 5 It is a structural block diagram of a data dimension reduction processing device in one embodiment;

[0058] Figure 6 FIG. 4 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0059] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0060] The data dimensionality reduction processing method provided in this application can be applied to Figure 1In the application environment shown. Among them, the terminal 102 communicates with the server 104 through the network. Among them, the user can determine to perform data processing by triggering the terminal 102. After receiving the instruction of data processing sent by the terminal 102, the server 104 extracts the historical analysis model from the database and parses it to obtain the initial feature. The server 104 can select the target feature according to the frequency of the acquired initial feature, so as to extract the initial data of the multidimensional data based on the selected target feature to obtain the initial data. Further, in order to facilitate the subsequent data processing, the server 104 can also perform data processing and dimensionality reduction processing on the initial data at the same data level after obtaining the initial data, and obtain the target data of the preset dimension, thereby reducing the amount of data for subsequent data processing and improving the processing efficiency of data processing. Among them, the terminal 102 can be, but is not limited to, various personal computers, laptops, smart phones, tablet computers and portable wearable devices, and the server 104 can be implemented with an independent server or a server cluster composed of multiple servers.

[0061] In one embodiment, Figure 2 As shown, a data dimension reduction processing method is provided, which is applied to Figure 1 The server in the example is used to illustrate the following steps:

[0062] Step 202: extracting a historical analysis model from a database, and parsing the historical analysis model to obtain an initial feature having a priority greater than a target priority.

[0063] The database may be a database of a server. The historical analysis model may be a model pre-configured in the server database. The historical analysis model configured in the server database may include models corresponding to different data types and related to different business types, including but not limited to models related to medical insurance, models related to disease types, etc., for example, a diabetes surgery analysis model, a routine examination analysis model, etc.

[0064] Specifically, the server can filter the historical analysis models stored in the database according to the filtering conditions, and then extract the filtered historical analysis models. For example, the server uses the medical insurance field as the filtering condition to filter and extract the historical analysis models related to medical insurance from the database.

[0065] Initial features refer to features obtained after parsing the historical analysis model, which may include but are not limited to cost features, frequency features, various indicator features, etc. Those skilled in the art will understand that the initial features described here refer to the meaning of the features and do not involve specific feature data. Specifically, cost features may include but are not limited to surgery fees, drug fees, examination fees, etc.; frequency features may include but are not limited to the number of medical visits, number of examinations, number of surgeries, number of drug purchases, etc.; various indicator features may include but are not limited to height and weight, heart rate, blood pressure, hemoglobin content, platelet count, glucose content, urine protein, etc.

[0066] The target priority level is a priority level preset by the server, which can be a high priority, a low priority, etc., or can be a primary, secondary, tertiary, etc. The target priority can vary according to different data types or business types, and this application does not impose any restrictions on this.

[0067] The priority of the initial feature can be associated with the historical analysis model. For different historical analysis models, the initial features extracted by the server with a priority greater than the target priority are different. For example, when the server analyzes the diabetes surgery analysis model, the initial features with a priority greater than the target priority may be features such as surgery fee, number of surgeries, and glucose content. When the server analyzes the routine examination analysis model, the initial features with a priority greater than the target priority may be features such as height, weight, blood pressure, heart rate, and vision.

[0068] Step 204, obtaining the frequency of the initial feature, and selecting the initial feature whose frequency meets the requirement as the target feature.

[0069] The frequency of the initial feature may be the frequency at which the initial feature appears in different historical analysis models. Different initial features may have different frequencies of appearance in multiple historical analysis models. For example, surgery fees may appear in various historical analysis models with a high frequency, while glucose content may only appear in a diabetes surgery analysis model with a low frequency.

[0070] Specifically, the higher the frequency of the initial feature appearing in different historical analysis models, if the initial feature appears in all historical analysis models, the more important the initial feature can be determined. Therefore, the initial feature can be determined according to the frequency of the feature appearing in different historical analysis models.

[0071] The frequency meeting the requirement may be that the frequency of the initial feature appearing in the historical analysis model meets a certain threshold condition, or the frequency of the initial feature appearing in the historical analysis model is sorted, and the initial features that meet certain requirements are the initial features whose frequency meets the requirement, for example, the top 10 in the sorting, etc. Specifically, the server may select the initial features according to the frequency of the initial features appearing in the extracted historical analysis model to determine the target features.

[0072] Step 206: extracting initial data corresponding to the target feature from the multidimensional data.

[0073] Among them, multidimensional data can refer to all data stored in the database, and may include newly added data each time the data is changed and historical data before the change. For example, corresponding to the medical insurance data mentioned above, the initial data refers to the medical data stored under the user's name generated after the user sees a doctor, and may include historical medical data and current medical data, which may specifically include but are not limited to consultation location, consultation time, International Classification of Diseases (ICD), registration department, registration doctor information, registration fee, payment method, examination items, examination fee, condition description, medical advice, drug list, drug price, drug dosage, payment window, medicine collection window, whether to follow up, follow-up time, number of consultations and other data.

[0074] Specifically, the server can extract initial data from multidimensional data based on the selected target features, and the extracted initial data can be divided into multiple categories. For example, for medical insurance data, it can include but is not limited to the current medical expense data, the current medical ICD data, and historical medical data. Among them, the current medical expense data can include but is not limited to surgery fees, drug fees, examination fees, etc.; the current medical ICD data can include but is not limited to the cost of the current confirmed ICD, the average cost of the ICD, etc.; the historical medical data can include but is not limited to the number of local outpatient visits, the number of local hospitalizations, the number of outpatient visits in other places, the number of hospitalizations in other places, the proportion of local outpatient visits, the proportion of outpatient visits in other places, and other data.

[0075] Step 208: Perform data processing of the same data level on the different types of initial data to obtain standard data.

[0076] Specifically, the extracted initial data may have a large difference in data magnitude due to the different data types. For example, the drug cost is 500, while the total cost is 1,000,000, and the data magnitudes of the two are very different.

[0077] The server can process the initial data of different data levels by the data processing method of the same data level to obtain standard data of the same data level. For example, continuing the previous example, the drug cost and the total cost are processed at the same data level to obtain drug costs and total costs with data levels between 0 and 100, that is, the obtained standard drug cost is 0.05 and the standard total cost is 100.

[0078] Specifically, the data processing method for the same data level can be selected according to different data types or different data levels. For example, square root, square, cube, exponential, logarithm and other methods can be selected, and this application does not impose any restrictions on this.

[0079] Step 210, performing nonlinear dimensionality reduction processing on the standard data to obtain target data of preset dimensions.

[0080] The preset dimension may be a dimension that is preset by the user to the server through the terminal according to the needs of subsequent data processing, and the data volume of the target data of the preset dimension may be smaller than the data volume of the standard data.

[0081] The methods of nonlinear dimensionality reduction processing may include but are not limited to isometric feature mapping (Isomap), locally linear embedding (LLE), modified locally linear embedding (MLLE), Hessian Eigenmapping, spectral embedding (Spectral Embedding), local tangent space alignment algorithm (LTSA), multi-dimensional scaling (MDS), t-distributed stochastic neighbor embedding (t-SNE), etc.

[0082] In practical applications, linear dimensionality reduction methods may also be used, including but not limited to principal component analysis (PCA), kernel PCA, incremental PCA, etc.

[0083] Specifically, the server can use the clustering characteristics of data in the multi-dimensional Riemann space according to the above method to map the multi-dimensional standard data to a low dimension, for example, to 2 dimensions, to obtain the target data.

[0084] In the above data dimensionality reduction processing method, the target features are obtained through the historical analysis model, and then the initial data corresponding to the target features in the multidimensional data are extracted, and the data of the same data level are processed to obtain standard data, and the standard data is subjected to nonlinear dimensionality reduction processing to obtain the target data of the preset dimension. The generated target data is generated based on the multidimensional data and is associated with the multidimensional data, so that the characteristics of the multidimensional data can be maintained, and then subsequent data processing and analysis can be performed through the target data. Compared with data processing and analysis using multidimensional data, it can save resource consumption when the system performs data processing and analysis, and can improve data processing efficiency.

[0085] As mentioned above, there may be multiple historical analysis models, which may be models respectively related to different data types or different business types.

[0086] In one embodiment, parsing the historical analysis model to obtain the initial feature having a priority greater than the target priority may include: parsing each of the historical analysis models separately to obtain the initial feature corresponding to each of the historical analysis models having a priority greater than the target priority.

[0087] For example, the diabetes surgery analysis model and the routine examination analysis model may be analyzed separately to obtain initial features such as surgery fee, examination fee, glucose content, etc. whose priority of the diabetes surgery analysis model is greater than the target priority, and to obtain features such as height, weight, vision, examination fee, etc., whose priority of the routine examination analysis model is greater than the target priority.

[0088] refer to Figure 3 The process diagram of the target feature determination step shown in the figure, wherein the frequency of the initial feature is obtained, and the initial feature whose frequency meets the requirements is selected as the target feature, may include the following process steps:

[0089] Step S302: determining the feature priority and the frequency of occurrence of each of the initial features in each of the historical analysis models.

[0090] Among them, feature priority refers to the metric of the feature's corresponding target priority in the historical analysis model, which can be high, medium, low, etc. For example, features such as surgery fee and glucose content are at a high priority in the diabetes surgery analysis model, so their feature priority can be high, while features such as height and weight are at a low priority in the diabetes surgery analysis model, so their feature priority can be low.

[0091] Those skilled in the art will appreciate that this is only an example, and in a specific application, the feature priority may also be a numerical metric, for example, high priority corresponds to 10 points, low priority corresponds to 1 point, etc. This application does not impose any restrictions on this.

[0092] Step S304 , comprehensively calculating the feature priority and the frequency of occurrence of each of the initial features in each of the historical analysis models, and obtaining the relative priority level of each of the initial features in each of the historical analysis models.

[0093] For example, after obtaining the feature priority and frequency of occurrence of each initial feature, the server can perform comprehensive calculations on each initial feature according to the set weight of the feature priority and the weight of the feature frequency of occurrence to obtain the relative priority level of each initial feature.

[0094] Those skilled in the art will understand that this is only an example, and in actual applications, the comprehensive calculation of the feature priority and frequency of occurrence of each initial feature in each historical analysis model can also be simply added or subtracted, or calculated using square, square root, exponential and other calculation methods, and this application does not impose any restrictions on this.

[0095] The relative priority level may be a high, medium, or low priority level, or may be a priority level expressed in the form of a numerical score such as 60 points, 70 points, 80 points, or 100 points, or a priority level determined by a combination of the two.

[0096] Step S306, comparing the relative priority levels of the initial features in the historical analysis models, and determining the initial features whose relative priority levels meet the requirements as target features.

[0097] Specifically, after the server calculates the relative priority levels of each initial feature, it can determine the target feature by comparing the relative priority levels of each initial feature. For example, it can determine the initial feature with a high relative priority level score as the target feature, or determine the initial feature with a high relative priority level as the target feature.

[0098] By obtaining the feature priority of each initial feature in each different historical analysis model and the frequency of occurrence of each initial feature, a comprehensive calculation is performed on each initial feature to obtain the relative priority level of each initial feature, and the target feature is selected according to the relative priority level. The target feature determined in this way has a strong correlation with the multidimensional data, which can improve the accuracy of the initial data obtained based on the target feature extraction.

[0099] In one of the embodiments, after extracting the initial data corresponding to the target feature in the multidimensional data, the above-mentioned data dimensionality reduction processing method may also include: obtaining the data source of the initial data and the user address corresponding to the initial data, and dividing the historical data in the initial data into local data and remote data based on the data source and the user address.

[0100] As mentioned above, for medical insurance data, the initial data extracted from the multidimensional data may include but are not limited to the current medical expense data, the current medical ICD data, and historical medical data. Among them, the historical medical data may include but are not limited to the number of local outpatient visits, the number of local hospitalizations, the number of out-of-town outpatient visits, the number of out-of-town hospitalizations, the proportion of local outpatient visits, the proportion of out-of-town outpatient visits, and other data.

[0101] Specifically, the server can divide the local outpatient visits, local hospitalizations, non-local outpatient visits, non-local hospitalizations, the proportion of local outpatient visits, the proportion of non-local outpatient visits and other data in the historical data according to the data source of the initial data and the user address corresponding to the initial data, so as to obtain local data (local outpatient visits, local hospitalizations, the proportion of local outpatient visits, etc.) and non-local data (non-local outpatient visits, non-local hospitalizations, the proportion of non-local outpatient visits, etc.).

[0102] In combination with the embodiments described above, after the server divides the initial data into local data and remote data, it performs data processing of the same data level on the different types of initial data to obtain standard data, which may include: performing data processing of the same data level on the different types of local data to obtain local standard data; performing data processing of the same data level on the different types of remote data to obtain remote standard data.

[0103] Specifically, the server performs data processing of the same data magnitude on the local data and the remote data, and the obtained standard data may be data of the same data magnitude or different. For example, the obtained local standard data may be data between 0-10, and the obtained remote standard data may be data between 0-100, or both may be data between 0-100, and the present application does not impose any limitation on this. Optionally, the methods of performing data processing of the same magnitude on the local data and the remote data may be the same or different.

[0104] Further, performing nonlinear dimensionality reduction processing on the standard data to obtain target data of preset dimensions may include: performing nonlinear dimensionality reduction processing on the local standard data and the remote standard data respectively to obtain local target data and remote target data of preset dimensions.

[0105] Specifically, the server may select different nonlinear dimensionality reduction processing methods to perform dimensionality reduction processing on the local standard data and the remote standard data according to different characteristics of the local standard data and the remote standard data.

[0106] Optionally, the server performs nonlinear dimensionality reduction on local standard data and remote standard data, and the dimensions of the obtained local target data and remote target data can be the same or different, which may be specifically associated with the selected data dimensionality reduction method and the data dimension of subsequent data processing. This application does not impose any restrictions on this.

[0107] By dividing the initial data into local data and remote data, and performing data processing and nonlinear dimensionality reduction processing on the same data level respectively, two different categories of target data are obtained, which is beneficial for subsequent data processing of different categories according to different regions, and can improve the pertinence and accuracy of subsequent data processing.

[0108] Optionally, after obtaining the local data and the remote data, the server may continue to divide the remote medical treatment data, for example, according to the specific province or city, into Beijing data, Shanghai data, Guangzhou data, etc.

[0109] In actual applications, the server may first process the initial data at the same data level, divide the standard data into regions based on the data source and user address, and then perform data dimensionality reduction to obtain local target data and remote target data of preset dimensions.

[0110] In one of the embodiments, before performing data processing of the same data level on different types of initial data to obtain standard data, the above data processing may further include: dividing the initial data into cost data and frequency data according to the data type.

[0111] As mentioned above, for medical insurance data, the initial data may include operation fees, drug fees, examination fees, local outpatient visits, local hospitalizations, non-local outpatient visits, non-local hospitalizations, etc. Specifically, the server may classify operation fees, drug fees, examination fees, etc. data into cost data, and classify local outpatient visits, local hospitalizations, non-local outpatient visits, non-local hospitalizations, etc. data into frequency data.

[0112] Furthermore, performing data processing of the same data order on the initial data of different types to obtain standard data may include: respectively obtaining preset formulas corresponding to the expense data and frequency data; performing data processing of the same data order on expense data of different data orders according to the preset formula corresponding to the expense data and the corresponding standard order of magnitude to obtain standard data corresponding to the expense data; performing data processing of the same data order on frequency data of different data orders according to the preset formula corresponding to the frequency data and the corresponding standard order of magnitude to obtain standard data corresponding to the frequency data.

[0113] The preset formula is a function formula corresponding to the data processing method of the same data level as described above, for example, a square root formula, a square formula, a cubic formula, an exponential formula, a logarithmic formula, etc. For different types of data, the preset formulas may be the same or different. Alternatively, different types of data may be associated with different preset formulas for data processing.

[0114] Specifically, there is a large difference in the data magnitudes of frequency data and expense data. By processing the frequency data and expense data at the same data magnitude, more accurate standard data can be obtained, thereby improving the accuracy of subsequent data processing.

[0115] In one embodiment, after extracting the initial data corresponding to the target feature from the multidimensional data, the server may also perform integrity polarity detection on the extracted initial data. Specifically, FIG. Figure 4 The flowchart of the data missing detection step is shown in FIG. 1 . The above data dimensionality reduction processing method may further include the following steps:

[0116] Step S402: Perform data missing detection on data of remaining data types in the initial data according to data of at least one data type in the initial data.

[0117] For example, continuing with the medical insurance data, after the server extracts the initial data, it determines based on the initial data that the user has been ill three times. However, in the initial data obtained, the treatment cost data only includes two groups of data, 2000 and 500. It can be determined that the treatment cost data is missing through the number of illness data.

[0118] Step S404: when it is detected that the initial data is missing, data of the data type is filled with data of the same data type as the missing data.

[0119] Continuing with the previous example, after the server detects that the treatment cost data is missing, it can determine the filling data based on the two sets of data 2000 and 500 included in the treatment cost data, and perform data filling. For example, it can be to take the average value or median of the two sets of data 2000 and 500, and fill the treatment data according to the determined average value or median.

[0120] In practical applications, the server can also perform data evaluation based on the user's historical data in the initial data, and fill in the data based on the evaluation data obtained. Preferably, the data evaluation can also be based on the user's historical data combined with the data of the user with the same medical record to obtain the evaluation data.

[0121] By positively detecting the acquired initial data and filling the polarity data, the integrity of the acquired initial data can be improved, thereby improving the accuracy of subsequent data processing.

[0122] It should be understood that although Figure 2-4 The steps in the flowchart are shown in sequence as indicated by the arrows, but these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified in this document, there is no strict order restriction for the execution of these steps, and these steps can be executed in other orders. Moreover, Figure 2-4 At least part of the steps may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily executed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be executed in turn or alternately with other steps or at least part of the sub-steps or stages of other steps.

[0123] In one embodiment, Figure 5 As shown, a data dimension reduction processing device is provided, which may include: a parsing module 100, a target feature generation module 200, a data extraction module 300, a same data level data processing module 400, and a dimension reduction processing module 500, wherein:

[0124] The parsing module 100 is used to extract relevant historical analysis models from the database, and parse the historical analysis models to obtain initial features with a priority greater than a target priority.

[0125] The target feature generation module 200 is used to obtain the frequency of the initial feature and select the initial feature whose frequency meets the requirements as the target feature.

[0126] The data extraction module 300 is used to extract initial data corresponding to the target feature from the multi-dimensional data.

[0127] The same data level data processing module 400 is used to perform data processing of the same data level on different types of initial data to obtain standard data.

[0128] The dimension reduction processing module 500 is used to perform nonlinear dimension reduction processing on the standard data to obtain target data of a preset dimension.

[0129] In one embodiment, there are multiple historical analysis models, and the parsing module 100 is used to parse each of the historical analysis models respectively to obtain initial features corresponding to each of the historical analysis models whose priority is greater than the target priority.

[0130] The target feature generation module 200 may include:

[0131] The first determination submodule is used to determine the feature priority and the frequency of occurrence of the initial feature in the historical analysis model.

[0132] The calculation submodule is used to comprehensively calculate the feature priority and the frequency of occurrence of each of the initial features in each of the historical analysis models to obtain the relative priority level of each of the initial features in each of the historical analysis models.

[0133] The comparison and determination submodule is used to compare the relative priority levels of the initial features in the historical analysis models, and determine the initial features whose relative priority levels meet the requirements as target features.

[0134] In one embodiment, the data dimension reduction processing device may further include:

[0135] The data classification module is used to obtain the data source of the initial data and the user address corresponding to the initial data after the data extraction module 300 extracts the initial data corresponding to the target feature in the multidimensional data, and divide the historical data in the initial data into local data and remote data based on the data source and the user address.

[0136] The same data level data processing module 400 may include:

[0137] The first same data level data processing submodule is used to perform data processing of the same data level on different types of local data to obtain local standard data.

[0138] The second same data level data processing submodule is used to perform data processing of the same data level on different types of remote data to obtain remote standard data.

[0139] The dimension reduction processing module 500 is used to perform nonlinear dimension reduction processing on the local standard data and the remote standard data respectively to obtain local target data and remote target data of preset dimensions.

[0140] In one embodiment, the data dimension reduction processing device may further include:

[0141] The classification module is used to classify the initial data into expense data and frequency data according to the data type before the same data level data processing module 400 performs data processing of the same data level on the initial data of different types to obtain standard data.

[0142] The same data level data processing module 400 may include:

[0143] The acquisition submodule is used to respectively acquire the preset formulas corresponding to the expense data and the frequency data.

[0144] The third same data level data processing submodule is used to perform data processing of the same data level on expense data of different data levels according to a preset formula corresponding to the expense data and the corresponding standard level, so as to obtain standard data corresponding to the expense data.

[0145] The fourth same data level data processing submodule is used to perform data processing of the same data level on frequency data of different data levels according to a preset formula corresponding to the frequency data and the corresponding standard order of magnitude, so as to obtain standard data corresponding to the frequency data.

[0146] In one embodiment, the data dimension reduction processing device may further include:

[0147] The detection module is used to perform data missing detection on the data of the remaining data types in the initial data according to the data of at least one data type in the initial data after the data extraction module 300 extracts the initial data corresponding to the target feature in the multidimensional data.

[0148] The filling module is used to fill the data of the data type with data of the same data as the missing data when it is detected that the initial data has data missing.

[0149] For the specific definition of the data dimensionality reduction processing device, please refer to the definition of the data dimensionality reduction processing method above, which will not be repeated here. Each module in the above-mentioned data dimensionality reduction processing device can be implemented in whole or in part by software, hardware and a combination thereof. The above-mentioned modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory in the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above modules.

[0150] In one embodiment, a computer device is provided. The computer device may be a server, and its internal structure diagram may be as follows: Figure 6As shown. The computer device includes a processor, a memory, a network interface and a database connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store historical analysis model data and various data in the data processing process. The network interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a data dimensionality reduction processing method is implemented.

[0151] Those skilled in the art will understand that Figure 6 The structure shown in the figure is merely a block diagram of a portion of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. Specifically, the computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0152] In one embodiment, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor implements the following steps when executing the computer program: extracting relevant historical analysis models from a database, and parsing the historical analysis models to obtain initial features whose priorities are greater than target priorities; obtaining the frequencies of the initial features, and selecting initial features whose frequencies meet the requirements as target features; extracting initial data corresponding to the target features in multidimensional data; performing data processing of the same data order on different types of initial data to obtain standard data; and performing nonlinear dimensionality reduction processing on the standard data to obtain target data of preset dimensions.

[0153] In one embodiment, there are multiple historical analysis models implemented when the processor executes the computer program, and the parsing of the historical analysis models to obtain initial features with a priority greater than the target priority may include: parsing each of the historical analysis models separately to obtain initial features corresponding to each of the historical analysis models with a priority greater than the target priority. The obtaining of the frequency of the initial features and selecting the initial features whose frequencies meet the requirements as the target features may include: determining the feature priority and the frequency of occurrence of each of the initial features in each of the historical analysis models; comprehensively calculating the feature priority and the frequency of occurrence of each of the initial features in each of the historical analysis models to obtain the relative priority level of each of the initial features in each of the historical analysis models; comparing the relative priority levels of each of the initial features in each of the historical analysis models, and determining the initial features whose relative priority levels meet the requirements as the target features.

[0154] In one embodiment, after extracting the initial data corresponding to the target feature in the multidimensional data implemented by the processor executing the computer program, it can also include: obtaining the data source of the initial data and the user address corresponding to the initial data, and dividing the historical data in the initial data into local data and remote data based on the data source and the user address. The data processing of the different types of initial data at the same data level to obtain standard data includes: performing data processing of the different types of local data at the same data level to obtain local standard data; performing data processing of the different types of remote data at the same data level to obtain remote standard data. The nonlinear dimensionality reduction processing of the standard data to obtain the target data of the preset dimension includes: performing nonlinear dimensionality reduction processing on the local standard data and the remote standard data respectively to obtain local target data and remote target data of the preset dimension.

[0155] In one embodiment, before the processor executes the computer program to perform data processing of the same data magnitude on the different types of initial data to obtain standard data, the method may further include: dividing the initial data into expense data and frequency data according to the data type. The method may include: respectively obtaining preset formulas corresponding to the expense data and frequency data; performing data processing of the same data magnitude on the expense data of different data magnitudes according to the preset formula and the corresponding standard magnitude corresponding to the expense data to obtain standard data corresponding to the expense data; performing data processing of the same data magnitude on the frequency data of different data magnitudes according to the preset formula and the corresponding standard magnitude corresponding to the frequency data to obtain standard data corresponding to the frequency data.

[0156] In one embodiment, after extracting the initial data corresponding to the target feature from the multidimensional data implemented by the processor executing the computer program, it may also include: performing data missing detection on the data of the remaining data types in the initial data based on data of at least one data type in the initial data; when it is detected that the initial data has missing data, filling the data of the data type with data of the same data as the missing data type.

[0157] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the following steps are implemented: extracting relevant historical analysis models from a database, and parsing the historical analysis models to obtain initial features whose priorities are greater than target priorities; obtaining the frequencies of the initial features, and selecting the initial features whose frequencies meet the requirements as target features; extracting initial data corresponding to the target features in multidimensional data; performing data processing of the same data order on different types of initial data to obtain standard data; and performing nonlinear dimensionality reduction processing on the standard data to obtain target data of preset dimensions.

[0158] In one embodiment, when a computer program is executed by a processor, multiple historical analysis models are implemented, and the parsing of the historical analysis models to obtain initial features having a priority greater than a target priority may include: parsing each of the historical analysis models separately to obtain initial features corresponding to each of the historical analysis models having a priority greater than the target priority. The obtaining of the frequency of the initial features and selecting the initial features whose frequencies meet the requirements as the target features may include: determining the feature priority and the frequency of occurrence of each of the initial features in each of the historical analysis models; performing a comprehensive calculation of the feature priority and the frequency of occurrence of each of the initial features in each of the historical analysis models to obtain the relative priority level of each of the initial features in each of the historical analysis models; comparing the relative priority levels of each of the initial features in each of the historical analysis models, and determining the initial features whose relative priority levels meet the requirements as the target features.

[0159] In one embodiment, after extracting the initial data corresponding to the target feature from the multidimensional data implemented by the computer program when executed by the processor, it may also include: obtaining the data source of the initial data and the user address corresponding to the initial data, and dividing the historical data in the initial data into local data and remote data based on the data source and the user address. The data processing of the same data magnitude on the initial data of different types to obtain standard data includes: performing data processing of the same data magnitude on the local data of different types to obtain local standard data; performing data processing of the same data magnitude on the remote data of different types to obtain remote standard data. The nonlinear dimensionality reduction processing on the standard data to obtain the target data of the preset dimension includes: performing nonlinear dimensionality reduction processing on the local standard data and the remote standard data respectively to obtain the local target data and the remote target data of the preset dimension.

[0160] In one embodiment, before the computer program is executed by the processor and performs data processing of the same data magnitude on the different types of initial data to obtain standard data, it may also include: dividing the initial data into expense data and frequency data according to the data type. The data processing of the same data magnitude on the different types of initial data to obtain standard data may include: respectively obtaining preset formulas corresponding to the expense data and the frequency data; performing data processing of the same data magnitude on the expense data of different data magnitudes according to the preset formula and the corresponding standard magnitude corresponding to the expense data to obtain standard data corresponding to the expense data; performing data processing of the same data magnitude on the frequency data of different data magnitudes according to the preset formula and the corresponding standard magnitude corresponding to the frequency data to obtain standard data corresponding to the frequency data.

[0161] In one embodiment, after the computer program is executed by the processor to extract the initial data corresponding to the target feature in the multidimensional data, it can also include: performing data missing detection on the data of the remaining data types in the initial data based on data of at least one data type in the initial data; when it is detected that the initial data has missing data, filling the data of the data type with data of the same data as the missing data type.

[0162] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0163] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0164] The above-mentioned embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the invention patent. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the patent of the present application shall be subject to the attached claims.

Claims

1. A data dimension reduction processing method, the method comprising: Extracting a historical analysis model from a database, and parsing the historical analysis model to obtain an initial feature having a priority greater than a target priority, including: screening the historical analysis model stored in the database according to a screening condition, extracting the screened historical analysis model, parsing the historical analysis model to obtain an initial feature, the priority of the initial feature being greater than the target priority, the target priority being a preset priority level, and the initial feature referring to a feature meaning, and not involving specific feature data of the initial feature; Obtaining the frequency of the initial feature, and selecting the initial feature whose frequency meets the requirement as the target feature; Extracting initial data corresponding to the target feature from the multidimensional data; Performing data processing of the same data level on different types of initial data to obtain standard data; Performing nonlinear dimensionality reduction processing on the standard data to obtain target data of preset dimensions; The historical analysis model is multiple, and the initial features whose priority is greater than the target priority are obtained by parsing the historical analysis model, including: Analyze each of the historical analysis models respectively to obtain initial features corresponding to each of the historical analysis models whose priorities are greater than the target priority; The obtaining of the frequency of the initial feature and selecting the initial feature whose frequency meets the requirement as the target feature includes: Determining the feature priority and frequency of occurrence of each of the initial features in each of the historical analysis models; Comprehensively calculating the feature priority and the frequency of occurrence of each of the initial features in each of the historical analysis models to obtain the relative priority level of each of the initial features in each of the historical analysis models; Comparing the relative priority levels of the initial features in the historical analysis models, and determining the initial features whose relative priority levels meet the requirements as the target features; Before performing data processing of the same data level on the different types of initial data to obtain standard data, the method further includes: Dividing the initial data into cost data and frequency data according to data type; The performing data processing of the same data level on the different types of initial data to obtain standard data includes: Respectively obtain preset formulas corresponding to the cost data and the frequency data; According to the preset formula corresponding to the expense data and the corresponding standard order of magnitude, the expense data of different data orders are processed at the same data order to obtain standard data corresponding to the expense data; According to the preset formula corresponding to the frequency data and the corresponding standard order of magnitude, data processing of the same data order of magnitude is performed on the frequency data of different data orders to obtain standard data corresponding to the frequency data.

2. The method according to claim 1, characterized in that After extracting the initial data corresponding to the target feature from the multidimensional data, the method further includes: Acquire the data source of the initial data and the user address corresponding to the initial data, and divide the historical data in the initial data into local data and remote data based on the data source and the user address; The performing data processing of the same data level on the different types of initial data to obtain standard data includes: Performing data processing of the same data level on the different types of local data to obtain local standard data; Performing data processing of the same data level on different types of remote data to obtain remote standard data; The performing nonlinear dimensionality reduction processing on the standard data to obtain target data of preset dimensions includes: Nonlinear dimensionality reduction processing is performed on the local standard data and the remote standard data respectively to obtain local target data and remote target data of preset dimensions.

3. The method according to claim 1, characterized in that After extracting the initial data corresponding to the target feature from the multidimensional data, the method further includes: Performing data missing detection on data of remaining data types in the initial data according to data of at least one data type in the initial data; When it is detected that the initial data is missing data, data of the data type is filled with data of the same data type as the missing data.

4. A data dimension reduction processing device, characterized in that: The device comprises: A parsing module is used to extract a historical analysis model from a database, and parse the historical analysis model to obtain an initial feature whose priority is greater than a target priority, including: screening the historical analysis model stored in the database according to the screening condition, extracting the screened historical analysis model, parsing the historical analysis model to obtain the initial feature, the priority of the initial feature is greater than the target priority, the target priority is a preset priority level, and the initial feature refers to the feature meaning, and does not involve specific feature data of the initial feature; A target feature generation module is used to obtain the frequency of the initial feature and select the initial feature whose frequency meets the requirements as the target feature; A data extraction module, used to extract initial data corresponding to the target feature from the multidimensional data; A data processing module of the same data magnitude is used to perform data processing of the same data magnitude on different types of initial data to obtain standard data; A dimension reduction processing module, used for performing dimension reduction processing on the standard data to obtain target data of preset dimensions; There are multiple historical analysis models, and the parsing module is used to parse each of the historical analysis models respectively to obtain initial features corresponding to each of the historical analysis models whose priority is greater than the target priority; The target feature generation module comprises: A first determination submodule is used to determine the feature priority and the frequency of occurrence of the initial feature in the historical analysis model; A calculation submodule, used for comprehensively calculating the feature priority and the frequency of occurrence of each of the initial features in each of the historical analysis models, to obtain the relative priority level of each of the initial features in each of the historical analysis models; A comparison and determination submodule, used for comparing the relative priority levels of the initial features in the historical analysis models, and determining the initial features whose relative priority levels meet the requirements as the target features; The device also includes: A classification module, used for classifying the initial data into expense data and frequency data according to data type; The same data level data processing module comprises: An acquisition submodule, used to respectively acquire preset formulas corresponding to the expense data and the frequency data; A third same data level data processing submodule is used to perform same data level data processing on expense data of different data levels according to a preset formula corresponding to the expense data and a corresponding standard level, so as to obtain standard data corresponding to the expense data; The fourth same data level data processing submodule is used to perform data processing of the same data level on frequency data of different data levels according to a preset formula corresponding to the frequency data and the corresponding standard order of magnitude, so as to obtain standard data corresponding to the frequency data.

5. The device according to claim 4, characterized in that The device also includes: A data classification module, for obtaining the data source of the initial data and the user address corresponding to the initial data after the data extraction module extracts the initial data corresponding to the target feature in the multidimensional data, and dividing the historical medical treatment data in the initial data into local medical treatment data and non-local medical treatment data based on the data source and the user address; The same data level data processing module comprises: A first same data level data processing submodule is used to perform data processing of the same data level on different types of local medical treatment data to obtain local standard data; The second same data level data processing submodule is used to perform data processing of the same data level on different types of remote medical treatment data to obtain remote standard data; The dimension reduction processing module is used to perform dimension reduction processing on the local standard data and the remote standard data respectively to obtain local target data and remote target data of preset dimensions.

6. The device according to claim 4, characterized in that The device also includes: A detection module, configured to perform data missing detection on data of remaining data types in the initial data based on data of at least one data type in the initial data; The filling module is used to fill the data of the data type with data of the same data as the missing data when it is detected that the initial data has data missing.

7. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that: When the processor executes the computer program, the following steps are implemented: Extracting a historical analysis model from a database, and parsing the historical analysis model to obtain an initial feature having a priority greater than a target priority, including: screening the historical analysis model stored in the database according to a screening condition, extracting the screened historical analysis model, parsing the historical analysis model to obtain an initial feature, the priority of the initial feature being greater than the target priority, the target priority being a preset priority level, and the initial feature referring to a feature meaning, and not involving specific feature data of the initial feature; Obtaining the frequency of the initial feature, and selecting the initial feature whose frequency meets the requirement as the target feature; Extracting initial data corresponding to the target feature from the multidimensional data; Performing data processing of the same data level on different types of initial data to obtain standard data; Performing nonlinear dimensionality reduction processing on the standard data to obtain target data of preset dimensions; There are multiple historical analysis models, and the initial features whose priority is greater than the target priority obtained by parsing the historical analysis models when the processor executes the computer program include: Analyze each of the historical analysis models respectively to obtain initial features corresponding to each of the historical analysis models whose priorities are greater than the target priority; The obtaining of the frequency of the initial feature and selecting the initial feature whose frequency meets the requirement as the target feature, which is implemented when the processor executes the computer program, includes: Determining the feature priority and frequency of occurrence of each of the initial features in each of the historical analysis models; Comprehensively calculating the feature priority and the frequency of occurrence of each of the initial features in each of the historical analysis models to obtain the relative priority level of each of the initial features in each of the historical analysis models; Comparing the relative priority levels of the initial features in the historical analysis models, and determining the initial features whose relative priority levels meet the requirements as the target features; Before the processor performs data processing of the same data level on the different types of initial data to obtain standard data when executing the computer program, the process further includes: Dividing the initial data into cost data and frequency data according to data type; The step of performing data processing of the same data level on the different types of initial data to obtain standard data, which is implemented when the processor executes the computer program, includes: Respectively obtain preset formulas corresponding to the cost data and the frequency data; According to the preset formula corresponding to the expense data and the corresponding standard order of magnitude, the expense data of different data orders are processed at the same data order to obtain standard data corresponding to the expense data; According to the preset formula corresponding to the frequency data and the corresponding standard order of magnitude, data processing of the same data order of magnitude is performed on the frequency data of different data orders to obtain standard data corresponding to the frequency data.

8. The computer device according to claim 7, characterized in that After the extracting of initial data corresponding to the target feature from the multidimensional data implemented when the processor executes the computer program, the method further includes: Acquire the data source of the initial data and the user address corresponding to the initial data, and divide the historical data in the initial data into local data and remote data based on the data source and the user address; The performing data processing of the same data level on the different types of initial data to obtain standard data includes: Performing data processing of the same data level on the different types of local data to obtain local standard data; Performing data processing of the same data level on different types of remote data to obtain remote standard data; The performing nonlinear dimensionality reduction processing on the standard data to obtain target data of preset dimensions includes: Nonlinear dimensionality reduction processing is performed on the local standard data and the remote standard data respectively to obtain local target data and remote target data of preset dimensions.

9. The computer device according to claim 7, characterized in that After the extracting of initial data corresponding to the target feature from the multidimensional data implemented when the processor executes the computer program, the method further includes: Performing data missing detection on data of remaining data types in the initial data according to data of at least one data type in the initial data; When it is detected that the initial data is missing data, data of the data type is filled with data of the same data type as the missing data.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 3 are implemented.

Images