Pattern recognition is the process of taking in raw data, and applying actions on it, such as classifying the input data into classes based on key features. With this in mind, in this chapter's exercise, we want to be able to organize various key fields within the combine data into positional (position) groups. In other words, group player height, weight (and/or other physical characteristics of the athlete), as well as combine drill scores into position groups. Perhaps we can find the characteristics for a WR group?

Once we have our prediction displayed in Watson Analytics, go through following steps:

1. We can click on the View All option (near the upper right of the page shown next) to display the charts with the ranking of the top predictors and their respective predictive strength value:

2. Each predictive strength value is displayed in parentheses after each predictor. The strongest predictor is on top. We see that Watson Analytics has found the weight is the top predictor for position, followed by fortyyd, heightinchestotal, and bench:

3. To see the statistical details behind each predictor, you can click on a predictor chart. From the Main Insight page, you can then select to show or hide specific statistical details:

So, let's summarize. Focusing on the strongest positional predictors for the WR position, the perfect player prospect has an average weight of 201 pounds and just over 6 foot tall. Their performance in the 40-yard dash is 4.5, and the bench drill is 6.37.

Here are the results of our project, summarized using the Watson Analytics Assemble functionality:

In conclusion, we can see that what we assumed (height, weight, and speed being the most important attributes of a potential WR) is indeed of the most interest for a WR; this seems to match what Watson Analytics is telling us (at least based upon the given data)!