Model construction method and system

Abstract

The invention relates to a model construction method, a model construction system, a server and a non-temporary computer readable storage medium. The model construction method comprises the steps of providing an original database and a corresponding configuration file; configuring a feature acquisition unit based on a plurality of subunits according to the configuration file; acquiring sample features from the original database based on the configured feature acquisition unit; and constructing a deep learning model according to the plurality of sample features, thereby improving the reusability of a model construction system, improving the feature production efficiency, and effectively reducing the resource consumption.

Description

technical field [0001] The present application belongs to the field of computer software applications, in particular to a model building method and system. Background technique [0002] With the rapid development of Internet technology, human beings have entered the era of information explosion. The simultaneous presentation of massive amounts of information, on the one hand, makes it difficult for information acquirers to find the information they are interested in, and on the other hand, makes it impossible for ordinary users to obtain a large amount of information that few people care about, which seriously hinders information dissemination. [0003] In order to solve the above problems, a personalized recommendation system has emerged. The personalized recommendation system establishes the binary relationship between users and information, uses the existing selection process or similarity relationship to mine the personalized information that each user is potentially in...

AI Technical Summary

Problems solved by technology

[0007] 1) Each model includes a large number of features, and each feature needs to be obtained by the staff through multiple steps such as filtering, extraction, calculation, etc., resulting in the need for a large number of human resources and development cycles in each model construction process , reducing the development efficiency of model building

[0008] 2) These features that make up the model cannot be parameterized. For very similar A features and B features, because there is only a very small difference between A and B, most of the codes are consistent. Therefore, in order to distinguish A features from B features, which can only be distinguished by additional definitions or combinations, increasing the complexity of the system

[0009] 3) All the features of the construction model are scattered, and each has a different proportion in performance, so it is difficult to find the real hot spots on the system, which increases the complexity of system optimization and iteration, etc.

Images