Python Tutorial: Increasing successful detections using data resampling

Key Takeaways

in this video we're going to talk about data resampling methods they help us to better train our models to recognize for cases when there are only very few cases of fraud the most straightforward way to adjust the imbalance of your data is to understand Paloma jority class aka a known fraud cases or oversample the minority class aka the fraud cases with undersampling you take random draws from your known fraud observations to match the amount of fraud observations as seen on the picture with over sampling you take random draws from the fraud cases and copy those observations to increase the amount of fraud samples you have in your data both methods lead to having a perfect balance between fraud and on fraud data but there are drawbacks with random under sampling you are effectively throwing away a lot of data and information with over sampling you are simply copying data and therefore are training your model on a lot of duplicates let's see how you can implement these methods in practice you can implement resampling methods using pythons imbalanced learn module it is compatible with scikit-learn and allows you to implement these methods in just two lines of code as you can see here I import the package and take the random over sampler and assign it to method I simply fit the method on to my original feature set X and labels Y to obtain a resampled feature set X and resampled Y I plot the data sets here side by side such that you can see the effect of my resampling method the darker blue color of the data points reflect that there are more identical data points now the synthetic minority over sampling technique or ass moat is another way of adjusting the imbalance by over sampling your minority observations aka your fraud cases but with smote were not just copying the minority class instead as you can see in this picture smoke uses characteristics of nearest neighbors of fraud cases to create new synthetic for cases and thereby avoids duplicating observations you might wonder which one of these methods is the best well it depends very much on the situation if you have very large amounts of data and also many fraud cases you might find it computationally easier to understand whether than to increase data even more but in most cases throwing away data is not desirable when it comes to over sampling smoke is more sophisticated as it does not duplicate data but this only works well if your fraud cases are quite similar to each other if fraud is spread out over your data and not very distinct using nearest neighbors to create more fraud cases introduce a bit of noise in the data as the nearest neighbors might not necessarily be for cases one thing to keep in mind when using resampling methods is to only resample on your training set your goal is to better train your model by giving it balanced amounts of data your goal is not to predict your sin-debt examples always make sure your test data is free of duplicates or synthetic data such that you can test your model on real data only the way to do this is to first split the data into a train and test it as you can see here I then resample the training set only I fit my model into the resample training data and lastly I obtain my performance metrics by looking at my original not resembled test data these that should look familiar to you so