This repository contains codes to learn applied machine learning, focusing more on the techniques and methods than on the statistics behind these methods. The first two notebooks introduce the scikit learn toolkit through different data sets. The issue of dimensionality of data is discussed, and the task of clustering data, as well as evaluating those clusters, is tackled. Difference between a supervised (classification) and unsupervised (clustering) technique is described here. Different technique to be used for a particular dataset, engineer features to meet that need are also described.
Further, Supervised approaches for creating predictive models are described, and the scikit learn predictive modelling methods were used to understand process issues related to data generalizability (e.g. cross validation, overfitting).
The final two notebooks describe more advanced techniques, such as building ensembles, and practical limitations of predictive models. Two real world problems 1. Fraud detection using machine learning and 2. Understanding and predicting property maintainance fines were solved to apply the concepts learned from first two notebooks in real world scenerio.