Model reuse method, device, equipment and storage medium

Abstract

The present invention discloses a model reuse method, device, equipment and storage medium, the method comprising: receiving a model to be reused and a target data table, and obtaining a model reuse method; running the model to be reused, and having the model to be reused process the target data table based on the model reuse method; obtaining a processing result after the operation is successful, and outputting the processing result. Thus, through model reuse, the connection between the reused model and the original model is enhanced, and the synchronization and model reuse efficiency are improved.

Description

Technical Field

[0001] The present invention relates to the field of machine learning technology, and in particular to a model reuse method, device, equipment and storage medium. Background Art

[0002] With the development of computer technology, more and more technologies (big data, distributed, blockchain, artificial intelligence, etc.) are being applied in the financial field. The traditional financial industry is gradually transforming to financial technology (Fintech). However, due to the security and real-time requirements of the financial industry, higher requirements are also placed on technology.

[0003] In the process of machine learning modeling, there are often scenarios where model reuse is required. However, there is currently a lack of model reuse methods. The only way is to create a new model copy with fields and parameters consistent with the original model, and then introduce the new data table into the model copy to run it to get the results. The model copy has no strong connection with the original model, and if the original model is adjusted, the model copy needs to be re-established or adjusted. Therefore, it is difficult to synchronize the reused model with the original model, and the reuse efficiency is not high. Summary of the invention

[0004] The present invention provides a model reuse method, device, equipment and storage medium, aiming to enhance the connection between the reused model and the original model, improve the synchronization and model reuse efficiency.

[0005] To achieve the above object, the present invention provides a model reuse method, the method comprising:

[0006] Receive the model to be reused and the target data table, and obtain the model reuse method;

[0007] Running the model to be reused, and having the model to be reused process the target data table based on the model reuse method;

[0008] After the operation is successful, the processing result is obtained and output.

[0009] Preferably, the model reuse method includes same-component model transfer and same-component model prediction.

[0010] The step of processing the target data table by the model to be reused based on the model reuse mode includes:

[0011] If the model reuse mode is the same component model transfer, the target data table is processed according to the fields and parameters of the model to be reused;

[0012] If the model reuse method is same-component model prediction, the model to be reused predicts the target data table according to a built-in algorithm.

[0013] Preferably, the model reuse method includes inter-component prediction and inter-component cross-validation.

[0014] The step of processing the target data table by the model to be reused based on the model reuse mode includes:

[0015] If the model reuse mode is inter-component prediction, the model to be reused predicts the target data table according to a built-in algorithm;

[0016] If the model reuse method is cross-validation between components, the model to be reused cross-validates the target data table according to the built-in validation parameters and the custom validation parameters.

[0017] Preferably, before the step of receiving the model to be reused and the target data table, the step further includes:

[0018] Set up data input ports and model input ports;

[0019] The model to be reused is input through the model input port, and the target data table is input through the data input port.

[0020] Preferably, before the step of inputting the model to be reused through the model input port and inputting the target data table through the data table input port, the step further includes:

[0021] The first data table is trained and / or processed to obtain the model to be reused.

[0022] Preferably, before the step of inputting the model to be reused through the model input port and inputting the target data table through the data table input port, the step further includes:

[0023] The initial data table is split according to the splitting tool to obtain the target data table.

[0024] Preferably, the splitting tool comprises a data table input port and a target data table output port, and the step of splitting the initial data table according to the splitting tool to obtain the target data table comprises:

[0025] Inputting the initial data table into the splitting tool through the data table input port;

[0026] The initial data table is split according to the received screening conditions by the splitting tool, and the data meeting the screening conditions is saved as the target data table

[0027] In addition, to achieve the above-mentioned purpose, the present invention also provides a model reuse device, the model reuse device comprising:

[0028] The receiving module is used to receive the model to be reused and the target data table, and obtain the model reuse method;

[0029] A running module, used for running the model to be reused, and the model to be reused processes the target data table based on the model reuse mode;

[0030] The output module is used to obtain the processing result after the operation is successful and output the processing result.

[0031] In addition, to achieve the above-mentioned purpose, the present invention also provides a model reuse device, which includes a processor, a memory and a model reuse program stored in the memory. When the model reuse program is executed by the processor, the steps of the model reuse method as described above are implemented.

[0032] In addition, to achieve the above-mentioned purpose, the present invention also provides a computer storage medium, on which a model reuse program is stored, and when the model reuse program is executed by a processor, the steps of the model reuse method described above are implemented.

[0033] Compared with the prior art, the present invention provides a model reuse method, device, equipment and storage medium, which receive a model to be reused and a target data table, and obtain a model reuse method; run the model to be reused, and the model to be reused processes the target data table based on the model reuse method; obtain a processing result after the operation is successful, and output the processing result. Therefore, through model reuse, the connection between the reused model and the original model is enhanced, and the synchronization and model reuse efficiency are improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Figure 1 is a schematic diagram of the hardware structure of a model reuse device involved in each embodiment of the present invention;

[0035] Figure 2 It is a flowchart of the first embodiment of the model reuse method of the present invention;

[0036] Figure 3 This is a schematic diagram of a first scenario of an embodiment of a model reuse method of the present invention;

[0037] Figure 4 This is a schematic diagram of a second scenario of an embodiment of a model reuse method of the present invention;

[0038] Figure 5 is a flow chart of a second embodiment of the model reuse method of the present invention;

[0039] Figure 6 It is a functional module diagram of the first embodiment of the model reuse device of the present invention.

[0040] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0041] It should be understood that the specific embodiments described herein are only used to explain the present invention, and are not used to limit the present invention.

[0042] The model reuse device mainly involved in the embodiment of the present invention refers to a network connection device capable of realizing network connection, and the model reuse device may be a server, a cloud platform, etc. In addition, the mobile terminal involved in the embodiment of the present invention may be a mobile network device such as a mobile phone and a tablet computer.

[0043] Reference Figure 1 , Figure 1 It is a schematic diagram of the hardware structure of the model reuse device involved in each embodiment of the present invention. In the embodiment of the present invention, the model reuse device may include a processor 1001 (such as a central processing unit, CPU), a communication bus 1002, an input port 1003, an output port 1004, and a memory 1005. Among them, the communication bus 1002 is used to realize the connection and communication between these components; the input port 1003 is used for data input; the output port 1004 is used for data output, and the memory 1005 can be a high-speed RAM memory or a stable memory (non-volatile memory), such as a disk memory. The memory 1005 can optionally be a storage device independent of the aforementioned processor 1001. Those skilled in the art can understand that Figure 1 The hardware structure shown in the figure does not constitute a limitation of the present invention, and may include more or less components than those shown in the figure, or combine certain components, or arrange the components differently.

[0044] Continue to refer to Figure 1 , Figure 1 The memory 1005 as a readable storage medium may include an operating system, a network communication module, an application module, and a model reuse program. Figure 1 In the example, the network communication module is mainly used to connect to the server and perform data communication with the server; and the processor 1001 can call the model reuse program stored in the memory 1005 and perform the following operations:

[0045] Receive the model to be reused and the target data table, and obtain the model reuse method;

[0046] Running the model to be reused, and having the model to be reused process the target data table based on the model reuse method;

[0047] After the operation is successful, the processing result is obtained and output.

[0048] The processor 1001 may call the model reuse program stored in the memory 1005 and perform the following operations:

[0049] If the model reuse mode is the same component model transfer, the target data table is processed according to the fields and parameters of the model to be reused;

[0050] If the model reuse method is same-component model prediction, the model to be reused predicts the target data table according to a built-in algorithm.

[0051] The processor 1001 may call the model reuse program stored in the memory 1005 and perform the following operations:

[0052] If the model reuse mode is inter-component prediction, the model to be reused predicts the target data table according to a built-in algorithm;

[0053] If the model reuse method is cross-validation between components, the model to be reused cross-validates the target data table according to the built-in validation parameters and the custom validation parameters.

[0054] The processor 1001 may call the model reuse program stored in the memory 1005 and perform the following operations:

[0055] Set up data input ports and model input ports;

[0056] The model to be reused is input through the model input port, and the target data table is input through the data input port.

[0057] The processor 1001 may call the model reuse program stored in the memory 1005 and perform the following operations:

[0058] Train and / or process the first data table to obtain the model to be reused

[0059] The processor 1001 may call the model reuse program stored in the memory 1005 and perform the following operations:

[0060] The initial data table is split according to the splitting tool to obtain the target data table.

[0061] The processor 1001 may call the model reuse program stored in the memory 1005 and perform the following operations:

[0062] Inputting the initial data table into the splitting tool through the data table input port;

[0063] The initial data table is split according to the received screening conditions by the splitting tool, and the data meeting the screening conditions is saved as the target data table.

[0064] An embodiment of the present invention provides a model reuse method.

[0065] Reference Figure 2 , Figure 2 It is a flow chart of the first embodiment of the model reuse method of the present invention.

[0066] In this embodiment, the model reuse method is applied to a model reuse device, and the method includes:

[0067] Step S101: receiving a model to be reused and a target data table, and obtaining a model reuse method;

[0068] In this embodiment, the model to be reused includes a feature engineering model to be reused and a training algorithm model to be reused, wherein the feature engineering model to be reused refers to various feature engineering. Feature engineering includes feature extraction, feature selection, and feature construction, wherein feature extraction refers to the process of selecting a relevant feature subset from a given feature set; feature extraction is before feature selection, and the object it extracts is the original data, and the purpose is to automatically construct new features and convert the original data into a set of features with obvious physical meaning (such as geometric features, texture features) or statistical significance; feature construction refers to constructing new features from the original data. The feature engineering described in this embodiment includes feature missing value filling, feature normalization, feature clustering, feature binning, feature combination, feature transformation, feature encoding, etc.; in this embodiment, the training algorithm model to be reused refers to a model obtained according to various training algorithms, and the training algorithms include: binary classification, multi-classification, regression algorithm, clustering algorithm, recommendation algorithm, etc.

[0069] The model reuse mode is specified by the user as needed. For example, the user can specify the model reuse mode through instructions, codes, preset controls, etc. In addition, model reuse can also be set in the workflow diagram so that the model reuse process can be directly executed according to the workflow diagram. In this embodiment, the model reuse mode set by the user is saved in advance for subsequent calls. If the user does not specify the model reuse mode, a prompt that the model reuse mode is not specified is output, or the default model reuse mode is called.

[0070] Step S102: running the model to be reused, and having the model to be reused process the target data table based on the model reuse method;

[0071] According to the model to be reused, the target data table and the model reuse mode specified in the step S101, the target data table is read, and the model to be reused is run according to the model reuse mode.

[0072] Specifically, the model reuse method includes same-component model transfer and same-component model prediction, such as transferring a decision tree to a decision tree and transferring a prediction model to a prediction model.

[0073] The step S102: the step of processing the target data table by the model to be reused based on the model reuse mode includes:

[0074] Step S102a1: if the model reuse mode is same-component model transfer, the target data table is processed according to the fields and parameters of the model to be reused;

[0075] Generally, the model to be reused has fields and parameters, and the target data table can be processed based on the fields and parameters. Specifically, if the same component model transfer is to transfer a decision tree to a decision tree, the target data table is processed according to the fields and parameters of the decision tree to be reused to obtain the transferred decision tree. For example, if the decision tree x1 is used as the model to be reused, the target data table to be processed is processed according to the fields and parameters of the decision tree x to obtain the transferred decision tree x2.

[0076] Step S102a2: If the model reuse method is same-component model prediction, the model to be reused predicts the target data table according to a built-in algorithm.

[0077] It can be understood that the prediction component includes a prediction algorithm, and this embodiment marks the prediction algorithm of the model to be reused as the built-in algorithm. For the same component model prediction, the data in the target data table is predicted according to the built-in algorithm to obtain a prediction result.

[0078] In this embodiment, the same-component model transmission and the same-component model prediction both belong to the same-component model reuse. The same-component model reuse refers to reusing the parameters and methods of the model to be reused without resetting the parameters, and processing the input target data table according to the parameters and methods of the model to be reused. For example, if the model to be reused is a feature engineering model to be reused, the feature engineering model to be reused pre-processes the target data table according to the corresponding feature engineering algorithm to obtain a second feature engineering model; the second feature engineering model is marked as a processing result, and the second feature engineering model is output. For example, if the feature engineering algorithm is feature binning, the target data table is binned according to the fields and parameter configuration of the feature binning to obtain a feature binning model.

[0079] Specifically, refer to Figure 3 , Figure 3 1 is a schematic diagram of the first scenario of an embodiment of the model reuse method of the present invention. Figure 3 In the example, the model to be reused is feature bin 1. First, data table 1 is binned to obtain feature bin 1, which is the model to be reused. Feature bin 1 processes data table 2 according to its own fields and parameter configuration to obtain feature bin 2, which is the second feature engineering model. Thus, feature bin 1 is model reused to obtain feature bin 2, which greatly reduces the workload of modelers.

[0080] Furthermore, the model reuse method includes inter-component prediction and inter-component cross-validation.

[0081] The step S102: the step of processing the target data table by the model to be reused based on the model reuse mode includes:

[0082] Step S102b1: If the model reuse mode is inter-component prediction, the model to be reused predicts the target data table according to a built-in algorithm;

[0083] "Inter-component" refers to two or more models to be reused. For example, a model reuse process includes feature engineering, decision tree and other models at the same time. If the model reuse method is inter-component prediction, the model reuse method involves a prediction model. When the prediction model is predicted by the quantitative prediction method, the most important task is to establish a prediction mathematical model. A prediction model refers to the quantitative relationship between things described by mathematical language or formulas for prediction. It reveals the inherent regularity between things to a certain extent, and it is used as a direct basis for calculating the predicted value during prediction. The prediction model includes Kalman filter prediction model, neural network prediction model, combined prediction model, etc. If the model reuse method is inter-component prediction, the model to be reused predicts the target data table according to the built-in algorithm.

[0084] Step S102b2: If the model reuse method is cross-validation between components, the model to be reused cross-validates the target data table according to built-in validation parameters and user-defined validation parameters.

[0085] Cross-validation is a practical method for statistically cutting data samples into smaller subsets, and is mainly used in modeling applications, such as PCR and PLS regression modeling. Cross-validation refers to taking out most of the samples from a given modeling sample to build a model, leaving a small part of the samples to use the newly established model for prediction, and calculating the prediction error of this small part of the samples, and recording their square sum. In this embodiment, the inter-component prediction and inter-component cross-validation belong to inter-component model reuse. The inter-component model reuse means that the component to be reused and the obtained processing result are not the same model, and the model to be reused can be used as the model input of the prediction component and the cross-validation component.

[0086] Specifically, refer to Figure 4 , Figure 4 It is a schematic diagram of the second scenario of an embodiment of the model reuse method of the present invention. Figure 4 The preset operation type in a is the same component model prediction. The data table a1 is trained according to the preset training algorithm to obtain the decision tree a1. The decision tree a1 is the model to be reused. The decision tree a1 is input through the preset model input port, and the target data table: data table a2 is input through the preset data input port. The tree model of the decision tree a1 is directly used to process the data table a2. The classification result of the data table a2 can be obtained without training.

[0087] Continue to refer to Figure 4 , Figure 4 In b, the model to be reused is feature engineering. According to the fields and parameters of the feature engineering algorithm, the data table b1 is selected to obtain feature selection b1, and the feature selection b1 is the model to be reused; the feature selection b1 is trained to obtain a decision tree b1, and then the decision tree b1 is used to predict the target data table b2 to obtain the processing result prediction b1. Alternatively, verification is performed according to the decision tree b1 to obtain a cross-validation result: cross-validation b. In this way, reuse between different models is achieved.

[0088] Furthermore, if the model to be reused is a training algorithm model to be reused, the training algorithm model output by the training algorithm model component to be reused can be used as the model input of the prediction component and the cross-validation component. Both the prediction component and the cross-validation component are used to evaluate the performance of the training algorithm model on new data, so the input model must be highly consistent with the algorithm model. In this way, the model output of the decision tree is used as the model input of the cross-validation component to cross-validate the decision tree model training; the model output of the decision tree is used as the input of the prediction component to predict new data.

[0089] Step S103: After the operation is successful, a processing result is obtained and the processing result is output.

[0090] After the model to be reused runs successfully, the processing result can be obtained and output, wherein the processing result includes the second training model, prediction result, verification result, etc. It can be understood that according to the function of the model to be reused, the processing result corresponding to the model function can be obtained. For example, if the model reuse method is that the same component model transfer is to transfer the decision tree to the decision tree, then the decision tree is output; if the model reuse method is inter-component prediction, then the prediction result is output.

[0091] Through the above scheme, this embodiment receives the model to be reused and the target data table, and obtains the model reuse method; runs the model to be reused, and the model to be reused processes the target data table based on the model reuse method; obtains the processing result after the operation is successful, and outputs the processing result. Therefore, through model reuse, the connection between the reused model and the original model is enhanced, and the synchronization and model reuse efficiency are improved.

[0092] like Figure 5 As shown, the second embodiment of the present invention proposes a model reuse method based on the above Figure 2 In the first embodiment shown, before the step of inputting the model to be reused through the model input port and inputting the target data table through the data table input port, the step further includes:

[0093] Step S1001, setting a data input port and a model input port;

[0094] Specifically, the data input port and the model input port are set in the machine learning reuse component, the data input port is used for inputting data and / or data tables, and the model input port is used for inputting the model to be reused.

[0095] Furthermore, the input port can be set by setting the shape of the port, or marking the port for distinction. The shape of the data input port is set, for example, a circle, a diamond, a triangle, etc.; the shape of the model input port is set, for example, a hexagon, an octagon, a star, etc. Generally, in order to facilitate distinction, in the same model reuse method, the shape of the data input port and the shape of the model input port are different shapes.

[0096] Among them, the data input port is used for data input, and the data input port is connected to the data port; the model input port is used for model input, and the model input port is connected to the model port.

[0097] Furthermore, a model output port is set, and the model output port is used to output the processing result.

[0098] Step S1002, inputting the model to be reused through the model input port, and inputting the target data table through the data table input port;

[0099] The model to be reused is input through the preset model input port, and the target data table is input through the data table input port.

[0100] It can be understood that before the step of inputting the model to be reused through the model input port and inputting the target data table through the data table input port, the following steps are also included:

[0101] The first data table is trained and / or processed to obtain the model to be reused.

[0102] Specifically, the first data table may be trained according to a preset machine learning algorithm to obtain the model to be reused; or the first data table may be processed according to a preset feature engineering to obtain the model to be reused.

[0103] Furthermore, before the step of inputting the model to be reused through the model input port and inputting the target data table through the data table input port, the step also includes: step S1002-1 splitting the initial data table according to a splitting tool to obtain the target data table.

[0104] In the process of machine learning modeling, for different business scenarios, modelers need to split the target data table that can be used for modeling from the initial data table according to certain conditions. For example, in a scenario of personalized product recommendation, to perform recommendation modeling for a user scenario targeting middle-aged customers, modelers need to first split the target data table of 40-65 years old and male gender from the initial data table containing customer information, and then further process the split target data table according to relevant dimensions, where the relevant dimensions include purchasing power, common purchasing methods, shopping preferences, etc.

[0105] Since business modelers may lack code development and debugging capabilities, using code to split qualified data sets in the computer field has a high threshold and is prone to errors. It is necessary to extract data through database operations or process data through Excel. It is difficult to cope with situations with large amounts of data and many combined conditions. It is time-consuming and labor-intensive, and the experience is poor. There is a lack of a direct human-computer interaction method to quickly perform conditional splitting processing.

[0106] In this embodiment, the initial data table is split by a splitting toolkit, the splitting tool includes a data table input port and a target data table output port, and the step S1002-1: the initial data table is split according to the splitting tool, and the step of obtaining the target data table includes:

[0107] Step S1002-1a: inputting the initial data table into the splitting tool through the data table input port;

[0108] In this embodiment, the data splitting tool can be an EXCEL data splitting expert, a CSV splitting tool, an SQL data splitting tool, etc.

[0109] Step S1002-1b: Split the initial data table according to the received screening conditions using the splitting tool, and save the data that meets the screening conditions as the target data table.

[0110] The screening conditions are set by the modeler according to the project requirements: select the fields to be added from the initial data table, and set the corresponding splitting conditions for each dimension of the field. The splitting conditions are determined by operators and values. When the conditions involve multiple expressions, the expressions can be added or deleted through the "+" and "×" after the value input box, and each expression combines the conditions in an "and" relationship. In this embodiment, the operators include ">", "=", "!=", "<", ">=", "<=", "in", "between", etc.; save the splitting conditions and run the component. After the component runs successfully, a running report is generated for viewing. The running report displays the splitting results of the conditional splitting. The original data that meets the splitting conditions will be split to the target data table output port for output, and the data set that does not meet the splitting conditions will be split to the non-target data table output port for output.

[0111] Through the above scheme, this embodiment sets a data input port and a model input port; inputs the model to be reused through the model input port, and inputs the target data table through the data table input port; receives the model to be reused and the target data table, and obtains the model reuse method; runs the model to be reused, and the model to be reused processes the target data table based on the model reuse method; obtains the processing result after the operation is successful, and outputs the processing result. Therefore, through model reuse, the connection between the reused model and the original model is enhanced, and the synchronization and model reuse efficiency are improved.

[0112] In addition, this embodiment also provides a model reuse device. Figure 6 , Figure 6 Schematic diagram of functional modules of the first embodiment of the model reuse device of the present invention.

[0113] In this embodiment, the model reuse device is a virtual device stored in Figure 1The memory 1005 of the model reuse device shown is used to implement all functions of the model reuse program: for receiving the model to be reused and the target data table, and obtaining the model reuse method; for running the model to be reused, and the model to be reused processes the target data table based on the model reuse method; for obtaining the processing result after successful operation, and outputting the processing result.

[0114] Specifically, the model reuse device includes:

[0115] The receiving module 10 is used to receive the model to be reused and the target data table, and obtain the model reuse method;

[0116] An operation module 20 is used to operate the model to be reused, and the model to be reused processes the target data table based on the model reuse mode;

[0117] The output module 30 is used to obtain the processing result after the operation is successful, and output the processing result.

[0118] Furthermore, the operation module includes:

[0119] A processing unit, configured to process the target data table according to the fields and parameters of the model to be reused if the model reuse mode is same-component model transfer;

[0120] The first prediction unit is used to predict the target data table according to the built-in algorithm by the model to be reused if the model reuse mode is the same component model prediction.

[0121] Furthermore, the operation module includes:

[0122] A second prediction unit is used for predicting the target data table by the model to be reused according to a built-in algorithm if the model reuse mode is inter-component prediction;

[0123] A cross-validation unit is used to, if the model reuse mode is cross-validation between components, cause the model to be reused to cross-validate the target data table according to built-in validation parameters and custom validation parameters.

[0124] Furthermore, the receiving module further includes:

[0125] A setting unit, used for setting a data input port and a model input port;

[0126] An input unit is used to input the model to be reused through the model input port, and to input the target data table through the data input port.

[0127] Furthermore, the input unit further includes:

[0128] The first obtaining subunit is used to train and / or process the first data table to obtain the model to be reused.

[0129] Furthermore, the input unit further includes:

[0130] The second obtaining subunit is used to split the initial data table according to the splitting tool to obtain the target data table.

[0131] Furthermore, the second obtaining subunit includes:

[0132] An input subunit, used for inputting the initial data table into the splitting tool through the data table input port;

[0133] The splitting subunit is used to split the initial data table according to the received screening conditions through the splitting tool, and save the data that meets the screening conditions as the target data table.

[0134] In addition, an embodiment of the present invention further provides a computer storage medium, on which a model reuse program is stored. When the model reuse program is executed by a processor, the steps of the model reuse method described above are implemented, which will not be repeated here.

[0135] Compared with the prior art, the present invention proposes a model reuse method, device, equipment and storage medium, which includes: receiving a model to be reused and a target data table, and obtaining a model reuse method; running the model to be reused, and the model to be reused processes the target data table based on the model reuse method; obtaining a processing result after the operation is successful, and outputting the processing result. Therefore, through model reuse, the connection between the reused model and the original model is enhanced, and the synchronization and model reuse efficiency are improved.

[0136] It should be noted that, in this article, the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or system 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, article or system. In the absence of further restrictions, an element defined by the sentence "comprises a ..." does not exclude the existence of other identical elements in the process, method, article or system including the element.

[0137] The serial numbers of the above embodiments of the present invention are only for description and do not represent the advantages or disadvantages of the embodiments.

[0138] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus a necessary general hardware platform, and of course by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) as described above, and includes a number of instructions for a terminal device to execute the methods described in each embodiment of the present invention.

[0139] The above description is only a preferred embodiment of the present invention, and does not limit the patent scope of the present invention. Any equivalent structure or process transformation made by using the contents of the present invention specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A model reuse method, characterized in that: The method comprises: Receive the model to be reused and the target data table, and obtain the model reuse method; Running the model to be reused, and having the model to be reused process the target data table based on the model reuse method; The model reuse method includes same-component model transfer and same-component model prediction. The step of processing the target data table by the model to be reused based on the model reuse mode includes: If the model reuse mode is the same component model transfer, the target data table is processed according to the fields and parameters of the model to be reused; If the model reuse mode is same-component model prediction, the model to be reused predicts the target data table according to the built-in algorithm; The model reuse method includes inter-component prediction and inter-component cross-validation. The step of processing the target data table by the model to be reused based on the model reuse mode includes: If the model reuse mode is inter-component prediction, the model to be reused predicts the target data table according to a built-in algorithm; If the model reuse mode is cross-validation between components, the model to be reused cross-validates the target data table according to built-in validation parameters and custom validation parameters; After the operation is successful, the processing result is obtained and output.

2. The method according to claim 1, characterized in that The step of receiving the model to be reused and the target data table also includes: Set up data input ports and model input ports; The model to be reused is input through the model input port, and the target data table is input through the data input port.

3. The method according to claim 2, characterized in that Before the step of inputting the model to be reused through the model input port and inputting the target data table through the data input port, the following steps are also included: The first data table is trained and / or processed to obtain the model to be reused.

4. The method according to claim 2, characterized in that: Before the step of inputting the model to be reused through the model input port and inputting the target data table through the data input port, the following steps are also included: The initial data table is split according to the splitting tool to obtain the target data table.

5. The method according to claim 4, characterized in that The splitting tool includes a data table input port and a target data table output port. The step of splitting the initial data table according to the splitting tool to obtain the target data table includes: Inputting the initial data table into the splitting tool through the data table input port; The initial data table is split according to the received screening conditions by the splitting tool, and the data meeting the screening conditions is saved as the target data table.

6. A model reuse device, characterized in that: The model reuse device comprises: The receiving module is used to receive the model to be reused and the target data table, and obtain the model reuse method; A running module, used for running the model to be reused, and the model to be reused processes the target data table based on the model reuse mode, wherein the model reuse mode includes same-component model transfer and same-component model prediction, and the running module is specifically used for, if the model reuse mode is same-component model transfer, then the target data table is processed according to the fields and parameters of the model to be reused; if the model reuse mode is same-component model prediction, then the model to be reused predicts the target data table according to a built-in algorithm, wherein the model reuse mode includes inter-component prediction and inter-component cross-validation, and the running module is specifically used for, if the model reuse mode is inter-component prediction, then the model to be reused predicts the target data table according to a built-in algorithm; if the model reuse mode is inter-component cross-validation, then the model to be reused cross-validates the target data table according to built-in verification parameters and custom verification parameters; The output module is used to obtain the processing result after the operation is successful and output the processing result.

7. A model reuse device, characterized in that: The model reuse device includes a processor, a memory, and a model reuse program stored in the memory. When the model reuse program is executed by the processor, the steps of the model reuse method according to any one of claims 1 to 5 are implemented.

8. A computer storage medium, characterized in that The computer storage medium stores a model reuse program, and when the model reuse program is executed by the processor, the steps of the model reuse method according to any one of claims 1 to 5 are implemented.

Images