The repository is a learning exercise to:
- Apply the fundamental concepts of machine learning from an available dataset
- Evaluate and interpret and justify our results and interpretation based on observed dataset in Jupyter notebook
The analysis is divided into four sections, saved in juypter notebooks in this repository
- Identifying the problem and Data Sources
- Exploratory Data Analysis
- Pre-Processing the Data
- Build model to predict whether breast cell tissue is malignant or Benign
Aim:Identify the types of information contained in our data set We'll used Python modules to import external data sets for the purpose of getting to know/familiarize ourself with the data to get a good grasp of the data and think about how to handle the data in different ways.
Aim: Explore the variables to assess how they relate to the response variable In this notebook, we'll get familiar with the data using data exploration and visualization techniques using python libraries (Pandas, matplotlib, seaborn. Familiarity with the data is important which will provide useful knowledge for data pre-processing)
Aim: Find the most predictive features of the data and filter it so it will enhance the predictive power of the analytics model. We'll use feature selection to reduce high-dimension data, feature extraction and transformation for dimensionality reduction. This is essential in preparing the data before predictive models are developed.
Aim: Construct predictive models to predict the diagnosis of a breast tumor. We'll construct a predictive model using SVM machine learning algorithm to predict the diagnosis of a breast tumor. The diagnosis of a breast tumor is a binary variable (benign or malignant). we'll also evaluate the model using confusion matrix the receiver operating curves (ROC), which are essential in assessing and interpreting the fitted model.
Aim: Construct predictive models to predict the diagnosis of a breast tumor. We'll aim to tune parameters of the SVM Classification model using scikit-learn.