waitipeter/TANZANIA-WATER-WELLS-PROJECT

Machine learning # TANZANIA-WATER-POINTS PREDICTION USING CLASSIFICATION MODELS For this project, I used different classification models to predict the status of different water points in Tanzania. There are three status groups to be predicted i.e - Functional, - Non-functional and - Functional Needs repair. This makes this problem a Ternary classification but for better classification I combined functional water points and functional needs repair as class 1 and non-funtional water points as class 0. ## Business Problem: An NGO focused on locating wells that needs repair. The Tanzanian government is looking to find patterns in non-functional wells to influence how new wells are built. As an employee of the NGO i have been tasked with presenting findings on how the Ministry of Water can identify wells that are non-functional or in need of repair, so so that they can dispatch teams to fix them ## Requirements This project requires that; Your Jupyter Notebook should demonstrate an iterative approach to modeling. This means that you begin with a basic model, evaluate it, and then provide justification for and proceed to a new model. After you finish refining your models, you should provide 1-3 paragraphs discussing your final model - this should include interpreting at least 3 important parameter estimates or statistics. ## Data used The original data was obtained from the DrivenData 'Pump it Up: Data Mining the Water Table' competition. Basically, there are 4 different data sets; submission format, training set, test set and train labels set which contains status of wells. With given training set and labels set, competitors are wanted to build predictive model and apply it to test set to determine status of the wells and submit. In this project, we used train set and train label set which have 59400 water points data with 40 features. ## Methodology 1.Understanding Data 2.Cleaning and Exploring Data 3.Preparing Data to Modeling 4.Finding Binary Model & Baseline 5.Ternary Targe …