In this chapter, we went through the steps for developing a prediction model when the response variable is of a numeric type. We started with a neural network model that had 201 parameters and then developed deep neural network models with over 7,000 parameters. You may have noticed that, in this chapter, we made use of comparatively deeper and more complex neural network models compared to the previous chapter, where we developed a classification model for the target variable that was of a categorical nature. In both Chapter 2, Deep Neural Networks for Multiclass Classification, and Chapter 3, Deep Neural Networks for Regression, we developed models based on data that was structured. In the next chapter, we move on to problems where the data type is unstructured. More specifically, we'll deal with the image type of data and go over the problem of image classification and recognition using deep neural network models.

In the next chapter, we will cover the steps required to develop an image recognition and prediction model using deep neural networks.