Shared-Nearest-Neighbours-Based-DBSCAN

Problem Statement:

Implement given algorithm in Scheme programming language for clustering data points given in a dataset D. Assume an ordering of D where ith data point has index i.

Parameters:

• K
• ε (≤ K)
• MinPts (≤ K)

Algorithm:

1. Sparsify Similarity Matrix
2. Construct the shared neighbor graph G
3. Identify core points
4. Form clusters using core points
5. Identify noise points
6. Assign border points to clusters

Input Format:

Each input would be given in a file in the following format. First line would contain five space separated parameters as follows.

• N D K ε MinPts
• N = Number of points in the dataset
• D = Dimensions of each point
• K, ε, MinPts are parameters for the algorithm
• This is followed by N lines each representing a data point of D dimensions.