search

anlucara / saferwalks-map

This repository contains the algorithms development in python (K-means, etc) and the map visualization.
0 stars 0 forks source link

Research about algorithms to find the nearest k based on coordinates. #3

Closed ancara-art closed 3 years ago

ancara-art commented 3 years ago

Some initial links: https://towardsdatascience.com/using-scikit-learns-binary-trees-to-efficiently-find-latitude-and-longitude-neighbors-909979bd929b

We used: K-means Hdbscan We need to research for one or two more algorithms as HDBSCAN seems to not be properly for our project idea. Those two algorithms could be: spectral clusterin, agglomeration clustering, and ward hierarchical clustering. (https://scikit-learn.org/stable/modules/clustering.html)

ancara-art commented 3 years ago

We need to research one or two more algorithms as HDBSCAN seems to give different cluster output not really clustering all the children, (it may leave points without cluster if they are found to be out of the core area?) We could use other algorithms that could be: spectral clustering, agglomeration clustering, and ward hierarchical clustering. (https://scikit-learn.org/stable/modules/clustering.html)

ancara-art commented 3 years ago

We decided to implement: spectral clustering, agglomeration clustering, and ward hierarchical clustering. (https://scikit-learn.org/stable/modules/clustering.html)