Machine Learning Project 2

Machine learning Project 2 Option A – CS-433 – EPFL<br> Fall 2018<br> *Authors:* Gianluca Mancini, Tullio Nutta and Tianchu Zhang<br> *Laboratory:* LESO-PB – Solar Energy and Building Physics Laboratory<br> *Supervisors:* Roberto Castello and Dasaraden Mauree

Directory tree

│   README.md
│
├───data
│   │   README.md
│   │
│   ├───average_data
│   └───zip_files
├───results
│   ├───baseline
│   │   └───images
│   ├───neural_network
│   ├───random_forest
│   ├───results_analysis
│   └───ridge_regression
│       └───images
├───scripts_data_extractor
│       data_extractor.ipynb
│       README.md
│       regression_mat_builder.ipynb
│
├───scripts_features_selection
│       feature_selection_stepwise.ipynb
│       helpers.py
│       README.md
│
└───scripts_regression
        baseline.ipynb
        feature_for_ridge_season.txt
        helpers.py
        neural_network.ipynb
        random_forest.ipynb
        README.md
        result_analysis.ipynb
        ridge_regression.ipynb

*IMPORTANT: All the scripts are set to run with the current folder tree architecture. The directory structure is self-contained. Any changes in the structure should be incorporated into the scripts in their respective paths definition.*

Overview

A part of `/data, in which raw, intermediate and final data for regression is saved, and /results, the other folders contain scripts. /scripts_data_extractor includes all the code necessary to convert the raw data into the matrix used in all the other scripts. /scripts_features_selection contains the code performing feature selection. Finally, the code of the different method is located in /scripts_regression`. All the scripts results are used for the conlcusions. Refer to single folder readme for more information.

Data download

All the required data to run the algorithms can be found at this link (download only). Additional useful information about the download can be found at the /data folder [README.md](./data/README.md).

Hardware requirements

Random forest and ridge regression require at least 32GB of RAM to run with the current settings. To run them with lower memory, follow the instructions in /scripts_regression folder [README.md](./scripts_regression/README.md). All the other scripts (neural network included) can operate with 4GB of RAM, but 8GB is advised.

Required packages

Make sure the following library are installed and updated. Use appropriate command for your environment (`pip or conda`) to change release or install them.

• numpy - 1.14.3
• pandas - 0.23.0
• scipy - 1.1.0
• matplotlib.pyplot - 2.2.2
• pytorch-cpu - 0.4.1
• scikit-learn - 0.20.0
• scikit-image - 0.13.1
  
  All the instruction concerning single scripts are contained in the READMEs in the subfolders.