olaabaza / tumor-cancer-prediction

All details in README.md

Project: Tumor Cancer Prediction.

The objective of the projects is to prepare you to apply different machine learning algorithms to real-world tasks. This will help you to increase your knowledge about the workflow of the machine learning tasks. You will learn how to clean your data, applying pre-processing, feature engineering, classification methods.

preprocessing techniques.

• **Check for missing value. **

• map the class label

Transform the class labels from their original string representation (M and B) into integers

Feature Standardization.

Use sklearn to scale and transform the data

Data analysis.

• DataFrame .describe()

Calculating some statistical data like percentile, mean and std of the numerical values of the Series or DataFrame.

Correlation Matrix .

Observation:

• The f2 and f4 feature have a strong positive correlation with f6,f8 and f9 feature;
• The f21 and f22 feature have a weak correlation with f24,f8 and f19 feature;

countplot

Observation:

Number of benign tumor data more than number of malignant tumor data

boxplot

Observation:

most of the values are usually higher in malignant than that of benign

Sizes of training and validation sets.

80% of the data for training and the remaining 20% for validation.

Hyperparameter tuning.

SVM

Hyperparameter:

• Kernel:

  • sigmoid

    accuracy score =0.945054945054945.
    

  • linear

    accuracy score =0.978021978021978.

  • rbf

    accuracy score =0.978021978021978.

• gamma:

  • 0.001

    accuracy score =0.9560439560439561.

  • 0.0001

    accuracy score =0.7362637362637363.

  • 0.01

    accuracy score =0.978021978021978.

Decision Tree

Hyperparameter:

• max_depth:

  • (None)

    Accuracy score =0.9340659340659341.

  • (2)

    Accuracy score =0.9560439560439561.

  • (4)

    Accuracy score =0.945054945054945

• min_samples_leaf:

  • (10)

    Accuracy score =0.9560439560439561.

  • (6)

    Accuracy score =0.967032967032967

  • (4)

    Accuracy score =0.9340659340659341

• xgboost

Hyperparameter:

• max_depth:

  • (3)

    Accuracy score =0.978021978021978

  • (2)

    Accuracy score =0.967032967032967

  • (4)

    Accuracy score =0.967032967032967

• learning_rate:

  • (0.05)

    Accuracy score =0.978021978021978.

  • (0.5)

    Accuracy score =0.967032967032967

  • (0.10)

    Accuracy score =0.978021978021978

Dimensionality Reduction.

• PCA

• SVM:

  • (0.90)

    Accuracy score =0.978021978021978.

  • (0.50)

    Accuracy score =0.9340659340659341

  • (25)

    Accuracy score =0.978021978021978

• Decision Tree:

  • (0.90)

    Accuracy score =0.9340659340659341

  • (25)

    Accuracy score =0.9340659340659341

  • (0.70)

    Accuracy score =0.9230769230769231.

• xgboost:

  • (0.90)

    Accuracy score =0.945054945054945

  • (0.50)

    Accuracy score =0.9340659340659341

  • (24)

    Accuracy score =0.9560439560439561

Training Time graph.

Testing Time graph.

Summary

We applied _Decision Tree,_XGBoosts and Support Vector Machine (SVM)

algorithms to the Tumor Cancer dataset.

• To predict whether the Tumor cancer is malignant or benign. • Compared the performance results of all the algorithms based on_

the accuracy values. and showed that XGBoosts classifier is the best among all in determining benign and malignant tumors.