Introduction
In this work I combine statistical analysis like K-means clustering (K-mean) and Principal Components Analysis (PCA) as a pre-processing step to find specific patterns related to each depositional facie, solving as much as possible the multi-class problem. Feature selection is important to define and quantify which well logging measurements contribute mainly to pattern recognition. Tuning hyperparameters from each classification algorithm gives the best possible performance based on data structure, avoiding complications like overfitting, underfitting and multi-class. Finally, model evaluation using thresholds metrics measure the performance of KNN and SVM. Proper data preparation is needed before using these algorithms. All these analyses will be done using Python programming language which contains most of the algorithms and statistical tools needed to perform a complete analysis.
The following work is splited in three diferent codes ordered as:
1.- Data Analysis and Feature selection.
2.- K-means & PCA.
3.- KNN and SVM classification.