K-Nearest Neighbors Approach for Predicting Drug Solubility in Cosolvent Mixtures

Published: 23 July 2025| Version 1 | DOI: 10.17632/8cmgv5v2js.1
Contributors:
,
,

Description

This notebook develops a KNN model to predict drug solubility based on its physicochemical properties and system conditions. The process includes data loading and cleaning, hyperparameter optimization of k, and performance evaluation of the final model.

Files

Steps to reproduce

The dataset DatasetSolubilityDrugV1, used for training the neural network, is available at the following link: https://data.mendeley.com/datasets/g686s23sy5/1. It should be cited as: Delgado, Daniel Ricardo; Vergara, Mateo; Cardenas Torres, Rossember Edén (2025), “Database of drug solubility in cosolvent mixtures at different temperatures.”, Mendeley Data, V1, https://doi.org/10.17632/g686s23sy5.1

Institutions

  • Universidad de America
  • Universidad Cooperativa de Colombia

Categories

Computer Science, Chemistry, Data Science, Machine Learning, Applied Computer Science

Licence