matanbt / clustering Goto Github PK

Implement the modified Gram-Schmidt algorithm.

Implement the QR Iteration Algorithm.

The matrix is defined as follows: W_ij = exp(-||x_i - x_j|| / 2) for each 2 points in the space.

Change it so we are compatible with the new demands from the code.

Create a script that will generate random points for our program.

Calculate the Eigengap heuristic, based on the first n/2 first sorted eigenvalues.

Calculate U, the matrix with only k dimensions (that we got from the eigengap heuristic), and then calculate T, the normalized matrix.

From the weighted matrix W, calculate the Laplacian:

1. Calculate the degree matrix D
2. Calculate the inverse matrix D^-0.5
3. Calculate the Laplacian: I - (D^-0.5) * W * (D^-0.5)

Write the remaining project tasks for the tasks.py file.

this seems to cause failure in the computation of the matrices (as we calculate the inverse of a big number and then the inverse the their sums etc.)
Might not be really a way to solve it, but it's interesting edge-case

Calculate the maximum capacity of the project, for both 2D points and 3D points.