Useful tools
In this section, we will see some of the tools that are used while creating machine learning models. Here, we will be using the scikit-learn package, but these are available in many other libraries too. The overall functioning and the purpose remain the same.
Preprocessing
Preprocessing the input as well as target labels in settings such as classification or regression is as important as the model itself. Some of the techniques used are explained as follows.
Normalization
For the ease of the model to learn proper parameters through a training set, it is highly essential to normalize the values in a small range, usually 0 to 1.
Noise
For making the system more robust, the input can also be added with small Gaussian noise. In the case of images as input, the noise corresponds to salt and pepper noise.
Postprocessing
In the case of classification, the output of the model is an array of probabilities. To compute the predicted label for the input, we use an index with the maximum value of...
