There are many implemented projects stored in different folders, I need to record their locations when I am looking for them.
-
Lime, Visualization, Visualizaed Feature Importance
-
Lime, Visualization, MultiClass
- https://marcotcr.github.io/lime/tutorials/Tutorial%20-%20continuous%20and%20categorical%20features.html
- All the features have to be int or float, but you can indicate categorical features
-
SHAP, Feature Importance, Visualization
- https://github.com/hanhanwu/Hanhan_Data_Science_Practice/blob/master/Better4Industry/Feature_Selection_Collection/try_shap_xgboost.ipynb
- It also supports multiclass now:
- How to generate multi-class output: shap/shap#367
- shap/shap#367
-
Multi-Class, Confusion Matrix
-
Simplified User Trained Search Engine - NN, Neural Network, Interactive
-
p-value, t-test to compare 2 lists of values
-
Find Optimal Thresholds of the Model Prediction
- https://github.com/hanhanwu/Hanhan_Applied_DataScience
- You can use whatever metrics you want to use, but by changing thresholds, it will show you which threshold will get the optimal metrics value
-
geckodriver errors
-
The definition of "Lift"
- In frequent pattern mining: https://github.com/hanhanwu/Hanhan_Data_Science_Practice/tree/master/Frequent_Pattern_Mining
lift = confidence/expected confidence
- In frequent pattern mining: https://github.com/hanhanwu/Hanhan_Data_Science_Practice/tree/master/Frequent_Pattern_Mining
-
Ployly API Key, PyOD
-
Multi Media
- How to read multi-media: https://www.analyticsvidhya.com/blog/2017/03/read-commonly-used-formats-using-python/?utm_source=feedburner&utm_medium=email&utm_campaign=Feed%3A+AnalyticsVidhya+%28Analytics+Vidhya%29
- PDF
PyPDF
library: https://github.com/mstamy2/PyPDF2/tree/master/Sample_Code
- Excel
- Python read, write Excel: https://github.com/hanhanwu/Basic_But_Useful/blob/master/python_excel.py
-
IPython JS display, python tricks
-
Virtual Environment, IPython Kernel
-
PCA, Dimensional Reduction, Reconstruction
- [Python] https://github.com/hanhanwu/Hanhan_Data_Science_Practice/blob/master/make_sense_dimension_reduction.ipynb
- How to make PCA dimensional reduction visualization makes sense
- How to keep original features based on PCA results
- [R] https://github.com/hanhanwu/Hanhan_Data_Science_Practice/blob/master/AI_Experiments/autoencoder_dimensional_reduction.Rmd
- PCA, autoencoder reconstruction error
- How to plot PCA/t-SNE converted 2D data with original labels: https://github.com/hanhanwu/Hanhan_COLAB_Experiemnts/blob/master/Try_DeepWalk.ipynb
- t-SNE vs PCA vs LDA: https://github.com/hanhanwu/Hanhan_Data_Science_Resources2
- Search for "t-SNE"
- [Python] https://github.com/hanhanwu/Hanhan_Data_Science_Practice/blob/master/make_sense_dimension_reduction.ipynb
-
For Industry Suggestions
- Overall 1: https://github.com/hanhanwu/Hanhan_Data_Science_Practice/tree/master/Better4Industry
- Overall 2: https://github.com/hanhanwu/Hanhan_Applied_DataScience
- Learning Notes: https://github.com/hanhanwu/Hanhan_Applied_DataScience/blob/master/Learning_Notes.md
- Experience Notes: https://github.com/hanhanwu/Hanhan_Data_Science_Practice/blob/master/Better4Industry/ExperienceNotes.md
- Feature Selection Collection: https://github.com/hanhanwu/Hanhan_Data_Science_Practice/tree/master/Better4Industry/Feature_Selection_Collection
- Model Evaluation: https://github.com/hanhanwu/Hanhan_Data_Science_Practice/tree/master/Better4Industry/Model_Evaluation
- Deployment Reminder: https://github.com/hanhanwu/Hanhan_Data_Science_Practice/blob/master/Better4Industry/Deployment_Experience.md
- Prototype Toolkit: https://github.com/hanhanwu/Hanhan_Applied_DataScience/tree/master/Prototype_Toolkit