Summary: Advanced Sequence Data Analysis

Get a recap of some of the topics discussed in this chapter.

We'll cover the following

In this chapter, we discussed several advanced techniques used for dealing with sequence data. Specifically, we discussed how such techniques can be used to model, identify, or predict players’ strategies, next actions, and churn.

Probabilistic level approaches

We introduced and showed the use of various techniques, including probabilistic methods using classical planning:

  • BNs and DBNs

  • HMMs and POHMMs

  • MLNs

  • MDP

  • RNNs

  • Deep RNNs

  • LSTM

It is important to note that all these techniques have their advantages and disadvantages. We outlined some of those in this chapter. In general, it's important for us to understand whether an interpretable model is necessary for the communication chain in our company or for our research. If it is, then that will eliminate the use of techniques like RNN or deep RNNs in favor of other more white-box approaches. If the prediction rate and accuracy are more important and we have tons of data, then deep RNNs are an option.

Get hands-on with 1200+ tech skills courses.