MLProcess
2019-09-16 12:04:29 28 举报AI智能生成
机器学习基本流程图
作者其他创作
大纲/内容
Introduction
Project background
Data source
Else
Define Problem
Supervise or unsupervise
Classification or regression
Data size
Library Loading
Common library
Visualiaztion library
Preprocessing library
Feature Engineering library
Model library
Data Loading
Load train and test
Define function to divide and combind train and test
Contact dataframe if necessary
Basic Info Showing for Train and Test
Row size, column size, column dtype with info()
Null data size for every features
Anomaly data size
Independent anomaly
Min and max
Aggregation anomaly
Outlier
Context anomaly
Outlier
Data Visualization
Distribution
Continuous feature<br>
Wave figure<br>
Line figure
Categorical feature<br>
Bar figure<br>
Relation
Continuous feature<br>
Scatter figure<br>
Categorical feature<br>
Bar figure
Data Preprocessing
Dealing with missing data
Dealing with anomaly data<br>
Continuous feature discretization<br>
Label encoding<br>
One-Hot encoding
Continuous feature data Transform<br>
Normalization<br>
Feature Engineering
Build new feature with old features and something else infomation<br>
Feature selection
Feature variance
Relevance with target feature<br>
Model and Blend
Single model
Multi model<br>
Blend models<br>
Simple weight to models<br>
Meta model to blend other models<br>
Some Little Trick
Correction factor
Single factor
Multi factor
Model stacking
Data resample<br>
Model Evaluate
Pipeline and Persistence
Summary and Look Back
0 条评论
下一页
为你推荐
查看更多
抱歉，暂无相关内容 
