plinkogame

Revising ML with JS -Simulation of popular game Plinko using the KNN algorithm.

Concepts covered:

1. Identifying data, assembling them (for dropPosition, bounciness, size, bucketLabel)

2. Deciding the type of output

3. Picking algorithm (K nearest neighbor)

4. Using model(s) to predict - classification

KNN implementation a. Drop the ball around the board, and record values b. For each observation, take the absolute value c. Sort results from least to great d. What are the top k records- common values e. What came frequently, and where will it goto

Improving Efficiency

1. Adjust parameters of analysis
2. Add more features
3. Change prediction point
4. Accept, there might not be a good co-relation

Some important functions:

Function knn - To implement KNN algorithm Function distance - Check the distance between multiple features Function splitDataset - To split data into testSet and trainingSet Function minMax - To achieve Normalisation

Utility Library used: Lodash