A Method of Linear Regression Model Building for Healthy Pupae Number Based on Common Characters Indicators

Abstract

In the process of silkworm breeding, if the healthy pupae number can be predicted by using silkworm's characteristics indicators, it will provide important reference value and guidance for the subsequent breeding and characteristics improvement. In this paper, the statistical analysis of silkworm characteristics data in the past 25 years is carried out, and the linear regression model and model establishment process for predicting the healthy pupae number with common characteristics and health indicators as independent variables are given. The significance test of F test and t test and the correlation analysis of between dependent variable and independent variables showed that there is a significant linear regression relationship between the predicted value of the healthy pupae number and the indicators of trait and health. The established linear regression model predicts the healthy pupae number. The prediction error is within 2%, which is in line with the actual breeding requirements. The results are accurate. The linear regression model has certain practicability.