Machine learning classification task optimization method and device for anomaly detection

Abstract

The invention provides a machine learning classification task optimization method and device for anomaly detection. The method according to one aspect of the present disclosure includes: collecting sample data for reflecting a state of a production process or a product to form a sample data set, each piece of sample data in the sample data set having a plurality of features; determining one or more candidate principal component sets for the sample data set; and for each of the one or more candidate principal component sets, executing the following operations: projecting a sample data set onto the candidate principal component set to obtain a projected sample data set, training each of a plurality of candidate machine learning classification models by using the projected sample data set, and evaluating a performance score for each of the trained candidate machine learning classification models; and selecting the candidate machine learning classification model with the highest performance score and the corresponding candidate principal component set as an optimal machine learning classification model and an optimal principal component set.

Description

technical field [0001] The present disclosure relates generally to information processing, and more particularly, to methods and apparatus for optimizing machine learning classification tasks for anomaly detection. Background technique [0002] As an important branch of artificial intelligence, machine learning has developed greatly in recent years with the improvement of computing power, the emergence of new algorithms and models, and the supply of massive data. Classification is one of the most commonly used task scenarios for machine learning techniques, including applications in industry. For example, the use of machine learning classification models has greatly improved the efficiency and automation of discrete control and process control operations, including predictive maintenance and diagnostics, which can not only provide early warning and protective guidance but also improve Increased machine / equipment uptime and reduced maintenance costs, thereby minimizing or ev...

AI Technical Summary

Problems solved by technology

In many practical scenarios, although it is easy to obtain various data as input to the machine learning classification model to perform classification tasks, the data dimension is too high or there are too many variables, that is, it contains too many features or attributes, which will not only reduce the Computational performance and introduces noise and thus leads to increased analysis complexity and reduced analytical power

Although methods such as principal component analysis can be used to automatically achieve dimensionality reduction of high-dimensional data, the performance of machine learning classification tasks performed using such dimensionality-reduced data still needs to be improved

Images