The data set obtained from Kaggle at https://www.kaggle.com/uciml/breast-cancer-wisconsin-data/data. This repository demonstrates some basic data analysis operations to be performed in python using numpy, patplotlib and pandas.
The jupyter notebook file can be found at pynb file.
Following table shows the different classification techniques used.
| Classification Method |
|---|
| Random Forest (Gini) |
| Random Forest (Entropy) |
| Multi Layer Perception (ReLU) |
| Multi Layer Perception (tanh) |
| Multi Layer Perception (sigmoid) |
| SVM NuSVC |
| SVM LinearSVC |
| SVM SVC Linear |
| SVM SVC RBF |
| Gaussian Naive Bayes |
| Decision Tree (Gini) |
| Decision Tree (Entropy) |
| Logistic Regression |
| KNN |
After analyzing the accuracy of all the models, the conclusion reached are:
- KNN with p=2 and num of neighbour at 5% of total data sets gives 98.2% accuracy
- Multi Layer Perceptron with tanh activation with 12 hidden layers gives 98.2% accuracy on Validation Set