In this chapter, you have learned about the most computationally intensive methods, SVMs and neural networks. Support vector machines perform really well on data in which the number of dimensions is very high, where other methods fail to work. By utilizing kernels, SVMs can reach very high test accuracies; we have covered how kernels actually work in detail in this chapter. Neural networks have become very popular in recent times for solving various problems; here, we covered all the deep learning fundamentals required for building a neural network model using both scikit-learn and Keras. In addition, results were compared between scikit-learn and Keras models to show apple-to-apple comparison. By utilizing deep learning, many new-generation artificial intelligence problems can be solved, whether it is text, voice, images, videos, and so on. In fact, deep learning itself has become a separate domain altogether.
In the next chapter, we will be looking at recommendation engines using...
