Are Outliers Always a Problem?
no...sometimes we need to treat them with respect
Outliers are mostly seen as not needed. We will see an example where this is not true.
Calibrate Your Classifier
you might wonder how you were doing without it all along
Would you buy oranges weighed on an uncalibrated weighing scale? Then why would you trust an uncalibrated classifier?
Longest Common Subsequence
...with an unconventional approach
This is a three way LCS solution using a ‘sparse’ data structure and an intuitive conceptual model to aid your understanding.
Roc Curve Step by Step
...with precision-recall curve thrown in
The ROC curve is an important metric to compare classifiers. Learn how to draw one step-by-step.
Linear Regression From Scratch
Don't just tow the line, move it!
Imagine you are buying a car and you want to know about its mileage. You don’t want to go for the user reviews or the company’s claim of mileage. The option you are left with is to predict the mileage all by yourself. So, if you are an interested data scientist, why not give it a try?
Mergesort Cutaway
Sorting an array can be done in different ways. It ranges from the simpler bubble sort to more complicated ones such as the merge sort.
Learning Curve
Imagine you have four classifiers with similar accuracies. Are they really similar? Plotting a learning curve might reveal a hidden side to these classifiers.
Understand the Confusion Matrix
Blow away the confusion
You’ve created a classification model and come across a new concept called confusion matrix. However tough it may seem, a classification model evaluation is not complete unless you add in your confusion matrix.
Conda Takes Forever
Anaconda taking forever to update? Here is what worked for me.
The Validation Curve
Tells you where to get off
Sometimes, you need to know when to stop training your model; this curve helps you decide. Let us see the validation curve applied to an Adaboost classifier use case.