varunkhajuria10/fys-stk4155

Repository for Applied Data Sciences and MAchine Learning from course at UiO.

FYS-STK4155

Repository for Applied Data Sciences and Machine Learning.

PROJECT 1:

project1 is a directory that contains the necessary files for the Project 1.

"project1" folder contains the python notebook project_1.ipynb for the Franke Function implementation and regression tests. Inside project_1 folder you will find another folder called "srtm" which have the "srtm.ipynb" python notebook for the regressions of the Digital Terrain Model data (.tif also included in the folder). Produced figures are also in the respective folder since they were produced locally.

To execute the python notebooks, please make use of jupyter lab, jupyter notebook or ipython notebook with support for Python version 3.6 or higher.

Necessary packages: pandas, numpy, matplotlib, mpl_toolkits, math, scipy, imageio, time, seaborn and sklearn

PROJECT 2:

project 2 is a directory that cointais the necessary files for the Project 2.

"project2" folder contains the following:

• Jupyter notebook "project2.ipynb" which runs all the studied features in this project. Is necessary of the Jupyter Notebook or Jupyter Lab to run it.
• nn.py, a python script where Neural Network and Logistic Regression classes are located.
• method.py, a python script that imports several useful functions as accuracy score and MSE calculations, plotting functions, Stochastic Gradient Descent methods, encoder and decoders of logist for Neural Networks, Design Matrix for Polynomial Regressions, and Franke Function data generation.

Necessary packages or Libraries. Before running, make sure you have the following packages: - numpy - math - random - sklearn - matplotlib - seaborn - tensorflow

Packages can be installed using the comand "pip install #name_of_the_package#" or "conda install #name_of_the_package#" in case you are using a conda environment.

PROJECT 3:

project3 is a directory that contains the necessary files for the Project 3.

"project3" folder contains the following:

• Jupyter notebook "nn_on_glaciers.ipynb" which runs all the studied features in this project. Is necessary of the Jupyter Notebook or Jupyter Lab to run it. It also as explanation of processes and validation steps to get to final outputs.
• nn.py, a python script where Neural Network (Beta NN) and U_Net classes are located (architecture). However this time the Beta NN model was not used, so you will find it declared inside the Jupyter Notebook file.
• method.py, a python script that imports several useful functions for data augmentation on images and pre-processing of landsat satelite images.

Necessary packages or Libraries. Before running, make sure you have the following packages: - numpy - math - random - sklearn - matplotlib - seaborn - tensorflow - rasterio

Packages can be installed using the comand "pip install #name_of_the_package#" or "conda install #name_of_the_package#" in case you are using a conda environment.