Hruday Chilukuri - 24JE0273
┃ Methods I implemented in algorithms made in the library ┃ ・Linear Regression : Calculation of Gradient Descent to Plot the Suitable Curve
・Polynomial Regression : Calculation of Gradient Descent to Plot the Suitable Curve
・Logistic Regression : Usage of Sigmoid Function with a Threshold of 0.5
・K- Nearest Neighbor : Usage of K-Nearest Neighbour with K = 3
・K - Means clustering : Euclidean Distance (Tried Manhattan Distance Also)
・Decision Trees : Implemented Decision Trees
・N-Neural Network : Implemented a simple 3 Layer Neural Network
┃ i experimented and played with:- ┃
・Linear Regression : Played with Epochs, Found best Learning Rate, Used MSE and R2 to find errors.
・Polynomial Regression : I changed the degree "N" of the equations, Learning Rate and Played with Epochs (Iterations)
・Logistic Regression : Usage of Sigmoid/Other Functions like Grad Descent, Epochs, Cross Entropy for best Accuracy
・K- Nearest Neighbors : Changed the Number of K with 2,3,4 and then Epochs and found F1 Score
・K Means Clustering : Used Euclidean/Manhattan/Minkowski Distances and changed between K's and found K = 2 to be best for given data.
・Decision Tree : Changed the Minimum Split, Maximum Depth, Used Gini rather than Entropy as it takes less time to process
・N Neural Network : Played with number of layers, like 3 Layer Neural Network, etc.
┃・Training Visualisations:- Attached in PDF and Code ┃