Page 1 of 1

KNN algorithm used for machine learning

Posted: Sun Dec 11, 2022 1:15 am
by togFox
I get that this is super niche and most of you won't use this but this is my favourite community and this might 'pay back' someone in the future.

K-nearest neighbour is an algorithm used by many machine learning applications to compare a single data-point against a largely populated data-set of similar values and then predicts how that new data-point should be categories.

For example, if you have historical sales information on the sex, age, post code and model of car purchased in the past, KNN can help predict what model of car future customers will lean towards. Another example: if you have a dataset of peoples height, weight and t-shirt size, KNN can help predict the what size t-shirt your next customer will likely purchase.

This library accepts a csv file with variable predictors and one label. It will return a prediction and the strength or confidence of that prediction.

There is no GUI or screenshots as it is a machine-learning algorith and not very sexy in that regard! This is my first ever public release of a module so happy for feedback and suggestions if anyone every uses it.

https://github.com/togfoxy/KNN-machine- ... /tree/main

Re: KNN algorithm used for machine learning

Posted: Mon Dec 12, 2022 6:13 pm
by GVovkiv
smart cookie