Supervised Learning - Classification Week 7 Challenge

1. Pros and cons of SVM.

Pros	Cons
- It works really well with a clear margin of separation	- It doesn’t perform well when we have large data set because the required training time is higher.
- It is effective in high dimensional spaces.	- It also doesn’t perform very well, when the data set has more noise i.e. target classes are overlapping.
- It is effective in cases where the number of dimensions is greater than the number of samples.	- SVM doesn’t directly provide probability estimates, these are calculated using an expensive five-fold cross-validation.
- It uses a subset of training points in the decision function (called support vectors), so it is also memory efficient.	- better understanding of hyper parameters is required to optimize the algorithm.

2. Apply SVM to the “Surface defects in stainless steel plates” dataset and evaluate it.

- labeled the dependent variables to a single dimentioanal array.

- droped TypeOfSteel_A400 column because its just the distinction from TypeOfSteel_A400 and TypeOfSteel_A300. this helps avoiding co_linearity among features and value 1 and 0 in TypeOfSteel_A300 is enough to differentiate the two.

- confusion matrix:

- tried different kernels and rbf returned high accuracy values.

program and data files are attached below.