In this chapter, we have seen how to develop an ML project using the RNN implementation, and called LSTM for HAR using the smartphones dataset. Our LSTM model has been able to classify the type of movement from six categories: walking, walking upstairs, walking downstairs, sitting, standing, and lying. In particular, we have achieved up to 94% accuracy. Later on, we discussed some possible ways to improve the accuracy further using GRU cell.

A convolutional neural network (CNN) is a type of feedforward neural network in which the connectivity pattern between its neurons is inspired by the animal visual cortex. Over the last few years, CNNs have demonstrated superhuman performance in complex visual tasks such as image search services, self-driving cars, automatic video classification, voice recognition, and natural language processing (NLP).

Considering these, in the next chapter we will see how to develop an end-to-end project for handling a multi-label (that is, each entity can belong to multiple classes) image classification problem using CNN based on the Scala and Deeplearning4j framework on real Yelp image datasets. We will also discuss some theoretical aspects of CNNs before getting started. Furthermore, we will discuss how to tune hyperparameters for better classification results.