Zero-based entry of data mining - transfer modeling to predict the parameters used car transaction price

0 Introduction

Zero-based entry of data mining - used car transaction price forecast of the title match with the data
zero-based entry of data mining - used car transaction price forecast of the title match to understand
zero-based entry of data mining - used car transaction price forecast of Data Analysis
zero Basics of Data Mining - forecast transaction price of second-hand car features engineering
paper stresses the various models expected to evaluate and adjust the reference model strategies.

1 learning content

1.1 linear regression model

• Linear regression requirements for the feature;
• Deal with long-tailed distribution;
• Understanding Linear regression model;
• Model works:
  • Linear Regression
  • Decision Tree
  • GBDT
  • XGBoost
  • LightGBM

1.2 Performance Verification Model:

• Evaluation function and objective function;
• Cross validation;
• Leaving a verification method;
• For verification time series problem;
• Drawing the learning curve;
• Curve drawing validation;

1.3 Embedded Feature selection:

• Lasso regression;
• Ridge regression;
• Decision tree;

1.4 Model Comparison:

• Common linear model;
• Common non-linear model;

1.5 model parameter adjustment:

• Greedy Scheduling method;
• Scheduling grid method;
• Bayesian parameter adjustment method;

1.6 recommended textbooks

• "Machine Learning"
• "Statistical learning methods"
• "Python war machine learning"
• "Machine learning-oriented features of the project"
• "Data scientists Interview"

Code, see my Github