Multi-class classification task
Task is to predict which water pumps are going to continue working, which are going to need repairs and which are going to fail.
(removed by Toogit admin)
Algorithms to use:
Regression, Decision Trees (Bagging, RF, XGBoost), Naive Bayes and KNN, KDE, PCA, LDA/QDA and SVM.
You must combine models by using ensembles. Using ensembles is about learning a target function by training a number of individual models and combining their predictions.
Python notebook must be used.
Feature Engineering - remove outliers, remove useless features, look for NAs and missing data and impute them, create new features
Any library but Scikit-Learn (Python) preferred.
!Reference Sources!
A detailed explanation of the machine learning process followed to achieve the results. Every cell - what is it looking for, why is this step carried out, what will it achieve, all results explained.
Minimum public score: 0.8262
Time to completion 48 hours after accepting job.
About the recuiterMember since May 20, 2018 Khoiril Anwar
from Laane-Viru, Estonia