ML-algos

A selection of Machine Learning algorithms which I have created, and may be applicable to other problems in the future

Convoluted neural network for classifying images of cats and dogs into their respective categories.

INCLUDES:

1. ConvertING 0-255 values into values between 0-1 for standardisation and reshaping images to aid with learning and prevent overfitting with small sample size,
2. flattening data to return a 1D array,
3. Sigmoid activation function in final dense layer to get probability output from CNN,
4. ADAM optimiser to combine Momentum and Root Mean Squared Propogation (RMSprop),
5. Binary crossentropy loss function useful for binary classification (i.e., cats or dogs, 1 or 0)

K-Nearest neighbours algorithm used for classification of books based upon rating given by different people.