Breast-Cancer-Detection

Breast Cancer Detection using Support Vector Machine

Overview:

Breast cancer is the most common cancer amongst women in the world. It accounts for 25% of all cancer cases, and affected over 2.1 million people in 2015 alone. It starts when cells in the breast begin to grow out of control. These cells usually form tumours that can be seen via X-ray or felt as lumps in the breast area.

Objective:

The key challenge against its detection is how to classify tumours into malignant (cancerous) or benign(non-cancerous). I will complete the analysis of classifying these tumours using machine learning (with SVMs) and the Breast Cancer Wisconsin (Diagnostic) Dataset. Since the labels in the data are discrete, the predication falls into two categories, (i.e., Malignant or benign). In machine learning, this is a classification problem. Thus, the goal is to classify whether the breast cancer is benign or malignant and predict the recurrence and non-recurrence of malignant cases after a certain period. To achieve this, we have used machine learning classification methods to fit a function that can predict the discrete class of new input.

Data Set Link:

https://www.kaggle.com/datasets/uciml/breast-cancer-wisconsin-data

About this dataset:

The collected data has been stored in the Comma Separated Value file data.csv. Each patient in the dataset is uniquely identified by their id. Broadly every tumour is classified into its dimensions, ‘radius’, ‘texture’, ‘perimeter’, ‘area’, ‘smoothness’, ‘compactness’, ‘concavity’, ‘concave points’, ‘symmetry’, ‘fractal dimension’. Each of these dimensions are then further divided into three more subcategories, ‘mean’, ‘se’ and ‘worst’ parameters. Every patient id contains the following variables:

• id: Unique id of every patient

• diagnosis: Whether the tumour is malignant(M) or benign(B)

• radius: Radius of the tumour (mean of distances from centre to points on the perimeter)

• texture: Texture of the tumour (standard deviation of grey-scale values)

• perimeter: Perimeter of the tumour (perimeter^2 / area - 1.0)

• area: Area of the tumour

• smoothness: Smoothness of the tumour (local variation in radius lengths)

• compactness: Compactness of the tumour

• concavity: Concavity of the tumour. (Severity of concave portions of the contour)

• concave points: Concave points of the tumour (number of concave portions of the contour)

• symmetry: Symmetry of the tumour

• fractal dimension: Fractal Dimension of the tumor ("coastline approximation" - 1)

Final Inferences

This work demonstrates the modelling of breast cancer as classification task using Support Vector Machine

The SVM performs better when the dataset is standardized so that all attributes have a mean value of zero and a standard deviation of one. We can calculate this from the entire training dataset and apply the same transform to the input attributes from the validation dataset.

1. We achieved an accuracy of 93% using the Support Vector Machine

2. After plotting the confusion matrix, we got the following information. • The classifier made a total of 174 predictions (i.e., 174 patients were being tested for the presence breast cancer).

   • Out of those 174 cases, the classifier predicted "yes" 43 times, and "no" 128 times.

   • In reality, 55 patients in the sample have the disease, and 116 patients do not.

3. We plotted the ROC - AUC Curve for our model and then found out the AUC value. It came out to be 0.97, which is good because our model has a good classification value.

4. I have tried to evaluate some other classifiers such as Decision Tree, Naive Bayes and SVM, using 10-fold cross validation technique. • KNN has 93.72% accuracy.

   • Decision Tree has 92% accuracy.

   • Naive Bayes has 93.71 % accuracy almost the same as KNN.

   • SVM has 90% accuracy.

5. KNN and Naive Bayes have the same accuracy as SVM which is around 93%.