Labor dispute prediction method based on Logistic regression algorithm

Abstract

The invention provides a labor dispute prediction method based on a Logistic regression algorithm. The labor dispute prediction method comprises the steps of 1, preprocessing labor dispute predictionrelated data; 2, constructing a regression function of the enterprise labor dispute occurrence probability based on the preprocessed labor dispute prediction related data, and constructing a likelihood function based on the regression function; step 3, solving an optimal parameter vector corresponding to the labor dispute prediction related data in the regression function based on the likelihood function so as to determine the regression function; and 4, calculating the labor dispute probability of each enterprise based on the determined regression function. According to the method, big data and advanced technology of machine learning are applied to enterprise labor dispute prediction, inaccuracy caused by human factors of traditional labor dispute prediction is fundamentally changed, andrecall ratio and precision ratio of model prediction are greatly improved.

Description

technical field [0001] The invention relates to the technical field of data prediction, in particular to a method for predicting labor disputes based on a Logistic regression algorithm. Background technique [0002] Labor disputes, also known as labor disputes, refer to disputes between laborers (employees) and investors (employers) due to various conflicts of interest. Labor disputes have been frequently occurring in recent years. Therefore, in order to nip possible labor disputes in the bud as much as possible and better maintain social stability, it is necessary to predict and warn enterprises that may have labor disputes. [0003] At present, most of the forecasts of enterprises that may cause labor disputes use manual scoring to grade all enterprises, and predict high-risk enterprises that are prone to labor disputes. Calculated as follows: [0004] y=a 1 x 1 +a 2 x 2 +a 3 x 3 +.....+a n x n [0005] where x 1 , x 2 , x 3 ,...,x n Scores for various indic...

AI Technical Summary

Problems solved by technology

[0006] With the development of the economy and more and more enterprises, the traditional labor dispute prediction method has the following two serious defects: First, because it uses manual scoring, it will waste a lot of human resources, and it is difficult to achieve real-time prediction ;Secondly, the weight coefficient of each scoring item is determined manually, and the score line for enterprises prone to labor disputes is also determined manually, resulting in low accuracy of prediction results

Images