moganad / machine-learning-on-breast-cancer-survival-prediction

We used different machine learning approaches to build models for detecting and visualizing important prognostic indicators of breast cancer survival rate. This repository contains R source codes for 5 steps which are, model evaluation, Random Forest further modelling, variable importance, decision tree and survival analysis. These can be a pipeline for researcher who are interested to conduct studies on survival prediction of any type of cancers using multi model data.

Machine-Learning-on-Breast-Cancer-Survival-Prediction

We used different machine learning approaches to build models for detecting and visualizing important prognostic indicators of breast cancer survival rate. This repository contains R source codes for 5 steps which are, model evaluation, Random Forest further modelling, variable importance, decision tree and survival analysis. These can be a pipeline for researcher who are interested to conduct studies on survival prediction of any type of cancers using multi model data.

The pipeline is as follows:

1. Model evaluation using 6 different algorithms in R (Model Evaluation.md)

1.1 Random Forest

1.2 Decision Tree

1.3 Support Vector Machine

1.4 Logistic Regression

1.5 Neural Networks

1.6 Extreme Gradient Boost

2. Random Forest Further modelling in R (Random Forest.md)

2.1 Selection of best ntree

2.2 Model evaluation for all the clusters

2.3 Calibration plot using Phyton 3

3. Variable Importance in R (Variable importance.md)

3.1 Using VSURF

3.2 Using randomForestExplainer

4. Decision Tree in R (Decision tree.md)

5. Survival analysis in R (Survival analysis.md)