# kmeans clustering fromscratch

## kmeans clustering - fromscratch

A numpy implementation of kmeans clustering

## Useful links

[Numpy indexing](https://numpy.org/doc/stable/user/basics.indexing.html)

## Code

```python
def kmeansclustering(obs:np.array, n_clusters:int, maxiter:int=100, maxflat:int=5, tol:float=0.01):
    np.random.seed(6)

    # initialise the centroids as random points selected from the dataset
    initial_centroids_i = np.random.choice(obs.shape[0], n_clusters, replace=False)  
    centroids = obs[initial_centroids_i,:]

    # initialise the distance matrix as an empty array with rows = tot obs and columns = n_clusters, 
    # distance_matrix[i][j] will be distance of obs i from cluster j
    distance_matrix = np.zeros((obs.shape[0], n_clusters))

    # keep track of the average distance between points and their cluster centroid
    # the stopping rule for the loop is a variation in avg. distance < tolerance level "tol" or a number
    # of iterations > maxiter
    prev_distance = np.inf
    flat = 0
    count = 0

    while flat < maxflat and count < maxiter:
        for c in range(n_clusters):
            # save distance from obs to each of the centroids into distance matrix
            distance_matrix[:, c] = np.linalg.norm(obs - centroids[c], axis=1)

        # the obs is assigned to the cluster represented by the nearest centroid
        clusters = np.argmin(distance_matrix, axis=1)

        # average distance between points and their cluster centroid
        avg_distance = np.mean(distance_matrix[range(len(distance_matrix)), clusters])

        # if variation < tol this iteration counts as a "flat" iteration
        if abs(prev_distance - avg_distance) < tol:
            flat += 1

        prev_distance = avg_distance

        for c in range(n_clusters):
            # compute the new centroids as clusters means
            cluster_obs = obs[np.where(clusters == c)[0], :]
            centroids[c] = np.mean(cluster_obs, axis=0)
    
    return clusters
```

## Github repository

<https://github.com/AndreaZoccatelli/kmeans_fromscratch>


---

# Agent Instructions: Querying This Documentation

If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter:

```
GET https://andrea-zoccatelli.gitbook.io/me/the-paper-vault/the-paper-vault/machine-learning/kmeans-clustering-fromscratch.md?ask=<question>
```

The question should be specific, self-contained, and written in natural language.
The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.