ML-Starter-Pack

1. Artificial Neural Network(ANN)

It contains an Artificial Neural Net implemented from scratch in Python.

2. KMeans

KMeans is an unsupervised Machine Learning Algorithm, which is implemented from scratch here. Iris dataset is used for testing and training

3. KNN

K- Nearest Neighbor Classification Algorithm is widely used for classification problem. Here, the algorithm is implemented from scratch and tested first on a small example of 2 dimensional test data and then on iris dataset.

4. Linear Regression

In Linear Regression, we try to establish relationship between 2 variables by fitting a line. The best fit line is known as regression line and it is represented by a linear equation Y = a*X +b This folder contains Linear Regression implemented from scratch in Python without using any Machine Learning libraries.

5. Logistic Regression

Logistic Regression is basically a classification algorithm which is used to estimate discrete values based on given data. The folder contains an ipython notebook depicting the implementation of logistic regression without any helper machine learning libraries.

6. Naive Bayes

Naive Bayes is a popular classification algorithm which assumes that one variable is independent of the other. In this example, the ipython notebook implements an email spam classifier.

7. Support Vector Machine(SVM)

This folder includes the following

1. An ipython notebook depicting basic implementation of SVM algorithm for learning basis. https://github.com/aribis369/ML-Starter-Pack/blob/master/SVM/SVM.ipynb

2. Optimized SVM using the SMO Algorithm https://github.com/aribis369/ML-Starter-Pack/blob/master/SVM/SVMl.ipynb

3. A pdf for explanation of SMO Algorithm https://github.com/aribis369/ML-Starter-Pack/blob/master/SVM/smo.pdf

The folder also contains .py files of the same and some datasets for illustrating the algorithms.

8. Anomaly Detector

Anomaly detection is a technique used to identify unusual patterns that do not conform to expected behavior, called outliers. It is used to detect intrusions in system, fraud detection and so on. The folder contains a python file which implements an anomaly detector from scratch.

9. Collaborative Filtering

Collaborative Filtering Algorithm is used in Recommender System. In collaborative filtering, the recommender system recommends items based on how users with similar interests liked it. The python file gives a collaborative filtering algorithm for recommending items to different users based on ratings and trust of reviews.

10. Dimension Reduction

PCA (Principal Component Analysis) algorithm is used for reducing dimensions. For example, If we want to reduce our data from 3D to 2D, we can use PCA algorithm. An ipython notebook as well as a python script is added in the folder which shows the implementation of this algorithm.

11. Examples implemented using Keras

1. Digit Recognition using MNIST dataset

12. Examples implemented using pyTorch

1. Classification of biomechanical orthopedic patients using Pytorch tensor
2. Implementing a neural network using pyTorch

13. Examples implemented using scikit

1. Analyzing Iris Dataset using Linear Regression, Logistic Regression, Decision Tree, Naive Bayes and SVM.
2. Analysis of Titanic dataset
3. Using Bernoullis Naive Bayes to classify a mushroom dataset and predict whether it is edible or not.

14. Examples implemented using Tensorflow

1. Digit classification from MNIST dataset using CNN

15. Gradient Descent

Implementation of gradient descent(an optimizing algorithm) from start.

16. Stock Prediction using Long Short-Term Memory (LSTM)

Recurrent Neural Networks allow information to persist, solving one major drawback of traditional neural nets. LSTMs are designed specially to avoid the long-term dependency problem. Remembering information for long periods of time is fortunately their default behavior, and not something they struggle to learn. A stock preditor using LSTM is implemented here.