The final motivation of this project is to build a Classifier to differentiate between old and young supernovae in realtime. This repository reflects all the efforts made towards achieving this final goal. It includes all the methods that were looked at and the various techniques that were learnt throughout the project. The repository has been organized in such a way that the outermost folders reflect different phases of the project while the inner folders are used to group scripts and images produced by different techniques.
The various folders in this repository consist of the following information.
-
EarlyColorData
The earliest part of the project where I learned about different classifiers like Neural Networks, Naive Bayes, Simple Vector Machines and Random Forests. I also learned about decision boundaries, confusion matrices and feature importances.
-
AdditionalTechniquesofAnalysis
Leanred about more techniques of data analysis like Prinicipal Component Analysis and Cross Validation
-
SimulatedDatasets
Started working with new datasets. Worked with KDE plots from the seaborn library. Looked at Dimenionsality reductions using t-SNE from sklearn as well as UMAP. Computed PCA loadings for the various datasets.