项目作者: khushi-411

项目描述 :
Conventional Algorithms in Machine Learning.
高级语言: Jupyter Notebook
项目地址: git://github.com/khushi-411/Machine-Learning-Algorithms.git
创建时间: 2021-02-28T07:22:27Z
项目社区:https://github.com/khushi-411/Machine-Learning-Algorithms

开源协议:MIT License

下载


Machine-Learning-Algorithms

This repository consists of some Conventional Machine Learning Algorithms. I used Social Network Ads Dataset for comparision between algorithms.

Data Visualization



20210409_194236.gif

Algorithms

Final Prediction Boundraries of each algorithm is shown bellow.

Linear Regression

lin1 lin2

Logistic Regression

log1 log2

SVM

Kernel Used: RBF

svm1 svm2

Decision Tree

Some more Visualization for Decision Tree:



20210409_225232.gif

dt3 dt4

Naive Bayes

nb1 nb2

KNN

k1 k2

Comparing Accuracys & Errors in Different Algorithms

  1. It includes (for both train set and test test:
  2. * Confusion Matrix
  3. * Accuracy
  4. * Mean Absolute Error
  5. * Mean Squared Error
  6. * Root Mean Sqaure Error
  7. * Precision
  8. * Recall
  9. * F-Score
  • For Training Data:

WhatsApp Image 2021-03-24 at 11 44 32 PM

  • For Testing Data:

WhatsApp Image 2021-03-24 at 11 44 32 PM (1)

Dependencies

  1. * Python: 3.7.10
  2. * Numpy: 1.19.5
  3. * Pandas: 1.1.5
  4. * Matplotlib: 3.2.2
  5. * Seaborn: 0.11.1
  6. * Bokeh: 2.3.0
  7. * Sklearn: 0.22.2.post1

License

It is provided with MIT License.