Deviation classification and parameter optimization method based on least square support vector machine technology

Abstract

The invention provides a deviation classification and parameter optimization method based on the least square support vector machine technology. A least square support vector machine is used for serving as a classifier, and is good in popularization capacity, and applicable to occasions with the requirement for high real-time performance, a virtual minority class oversampling algorithm is improved, influence of an isolating sample is eliminated, importance of a boundary sample is highlighted, classification can have a certain deviation, and the probability of wrongly classifying defective products into accepted products is reduced. According to the parameter optimization method based on the least square support vector machine technology, firstly, primary parameter optimization is carried out through a coupling simulated annealing algorithm, then sophisticated search is carried out by using a grid algorithm on the basis, the time for parameter optimization is reduced when a least square support vector machine model is trained, accuracy of classification is higher, and classification performance is improved.

Description

technical field [0001] The invention relates to the field of pattern recognition in machine vision detection, and specifically refers to a method for realizing biased classification of industrial products and a parameter optimization method of the least squares support vector machine through least squares support vector machine technology. Background technique [0002] Machine vision technology is an important branch of computer science. After more than 30 years of rapid development, with its advantages of fast speed, high precision and never fatigue, it gradually replaces manual visual inspection on industrial production lines, reducing the While reducing labor costs, strict control of product quality can be achieved. [0003] As an important field of machine vision, pattern recognition, a classifier with high classification accuracy and good real-time performance plays a vital role in the control of product qualification rate. In actual production, we should try our best ...

AI Technical Summary

Problems solved by technology

On the basis of the simulated annealing algorithm, the coupled simulated annealing algorithm considers several current states coupled together through energy, and has parallelism. While improving the optimal parameters, the convergence speed will not decrease, but it often needs to go through many times of annealing To find the optimal solution, time-consuming

Images