Comparison of Logistic Regression Model, Neural Network Model and Decision Tree Model on an Epidemiological Study
- Select Volume / Issue:
- Type of Publication:
- Logistic Regression Model, Neural Network Model, Decision Tree Model, Bovine Tuberculosis, Spearmans Rank Correlation, Lift Chart
- Renhao Jin; Fang Yan; Tao Liu
- This paper use three predictive models (logistic regression model, neural network model, and decision tree model) to an epidemiological study, i.e. bovine tuberculosis (bTB) occurrence in cattle herds, together with well-established risk factors in the area known as West Wicklow, in the east of Ireland. The binary target variable is whether the herd is in the restricted status, which is defined by whether any bTB reactor is detected in the herd. To compare the performance of the three model, the observations are divided into three part of Training data set (50%), Validation data set (30%), and Test data set (20%). The model performances are mainly based on the results of test dataset, and the decision tree model is the best model on this study. By analysis on the lift charts on test data set, the built decision tree model can be used to enhance practice efficiency.
Full text: IJISM_476_FINAL1.pdf