This article aims to shed light to the highly discussed yet confounding topic in machine learning: hyperparameter tuning. Up to this point, hyperparameter tuning is mostly talked about as a means to get the "best" model, with little being discussed about the how, the what and the why. Building machine learning models like a black …
Continue reading Comparing Grid Search and Randomized Search for Hyperparameter Tuning
Selecting the appropriate machine learning model for the data set requires knowledge and insights about the data acquired from the EDA stage. The clean data used for this exercise can be downloaded here and the post on how we cleaned this data is here. In our bivariate EDA, we learned that there isn't strong correlation …
Continue reading Machine Learning Model Building, Selection and Hyperparameter Tuning
The data set we use here is a simulated data that resembles closely to real-world data, and it can be downloaded here. It requires substantial data manipulation, including cleaning, binning, imputing and one-hot encoding before the machine learning model can train on it. Cleaning Thorough EDA in this post revealed several spelling mistakes, and we …
Continue reading Cleaning, Binning, Imputing and One-hot Encoding
Most data sets which are readily available, have been cleaned for pedagogical purposes. In this post, we present a case study with a realistic touch. This simulated data set has 40,000 entries and 100 predictors, riddled with spelling mistakes, missing entries, etc to mimic real-life data faced by data scientists. This exercise will give us …
Data Bay Today 