声明
摘要
Abstract
Content
Chapter 1.Introduction
1.1 Research significance and purpose
1.2 Research status and level
1.2.1 Current research status
1.2.2 Present Problems in the Field and Proposed Solution
1.3 Strueture of thesis
1.4 Summary
Chapter 2.Supervised Learning
2.1 Supervised learning
2.2 Main algorithms of supervised learning
2.2.1 Gradient Boosting
2.2.2 Random Forests
2.2.3 Extremely Randomized Trees
2.2.4 AdaBoost
2.2.5 Neural Networks
2.3 Naive Bayes classifier
2.4 R statistical programming language
2.5 Python
2.6 Summary
Chapter 3 Data Preprocessing
3.1 Use of Data Preprocessing Techniques
3.1.1 Instance selection and outliers detection
3.2.1 Feature Engineering
3.2.2 Preprocessing data
3.3 Visualization
3.3.1 Clusters of stations
3.3.2 Energy distribution by year
3.4 Summary
Chapter 4 Choosing the best algorithm in use validation
4.1 Use of three algorithms
4.2 Experiment process
4.2.1 Task description
4.2.2 The overall workflow
4.2.3 Comparison of 3 algorithms
4.2.4 Experiment results
4.3 Training and Test on a UCI dataset
4.4 Summary
Chapter 5 Conclusion and Future Works
References
ACKNOWLEDGEMENT