March Madness Series - Sweet Little Sixteen

Through a series of four challenges, you will build a tool that gathers information on teams and players by analyzing textual material from top sports writers and commentators- the material  could include game reports, columns, Twitter feed, blogs.  The tool should be able to help categorize players based on their skills, temperament, role/position etc, from these materials, using the capabilities of IBM Watson. 

This set of challenges will build on each other - so it will be of value to make each solution generic enough so that it could be reused in the next challenge. In the first challenge, we looked at teams with regard to Bracket predictions.
https://www.topcoder.com/challenges/30063503/

Description

In this second challenge, you are looking at players, teams and their positions and skills within the team with respect to the comments they have received from experts.

The task in this challenge is to identify some words and phrases that are important in the buildup or review of a game by expert commentators. Examples of positions in a team are guard, point guard, center, forward etc.  Skills could be rebounder, 3-point shooter etc. Use expert comments as the basis for finding relations between a player and his team, position and skills.

Select four colleges from the top 16. Take 10 to 15 documents per college and build your model using the attributes of the NLU system-  sentiment, category, concept, entities and relations. For the sake of uniformity across participants, the teams should be from the men’s tournament - no discrimination intended. Picking the strongest four teams will be an advantage.
 
Select some questions that with will be of interest to a Basketball analyst. The variables in the questions should be player position, skill and team.The questions you select will need to have at least one AND or OR operator in them. Examples:

• Name star performers in Kansas with 3-point shooting skills.
• Tell me about players who are good 3-point shooters or good rebounders.

You could look at  sites like  NCAA, Yahoo sports, ESPN, Sporting news etc.

In addition to your enriched model, you could also train the Discovery query service to return more relevant responses.
You may choose to build on your code from the first challenge - so that at the end of the four challenges, you will be able to demonstrate a more substantial application. The review for this challenge will focus only on the requirement of this challenge.

Requirements

1. Please join the Topcoder Cognitive Community if you have not already, and get an IBM Cloud Account by using this link.
2. Choose a few questions and show how your answers are a distinct improvement from the   answers using the default model. The questions you choose should include a conjunction or a disjunction in them.
3. The output answers should contain commentator name, publication, url of the comments/news item, a short extract from the comments.
4. Please also show the answers (with above mentioned details) using the default model, so that it will be easy to compare the results.
5. The questions should allow the user to enter values with regard to player skill, position and team.
6. A design document on your approach - why certain features you selected have led to better responses. 

Final Submission Guidelines

1. Deploy your application to your own IBM Cloud instance.
2. Upload a .zip containing your source code and a text file called ibm-cloud-deployment.txt.  This .txt file should contain the URL defined above for us to test.
3. You can use any programming language to build the application, as long it’s supported by IBM Cloud, has an api, provides a UI, and meets the spec criteria.
4. Detailed instructions on deploying and testing it locally. 

Review Guidelines

1. Richness of Model

• Does the model utilize and exploit NLU features?
• Quality and complexity of queries addressed by the model
• Richness of model will not score any points  if there is no implementation. However, the implementation could be for a section of the model.Implementation

2. Implementation

• IBM NLU and Discovery features used to demonstrate the solution
• Design and code quality

3. Documentation

• A document explaining your approach on how you have enhanced the default scores given by IBM Watson. Did anything not work you way you expected? How are you enhancing the output?
• A demo video of your solution

4. Ease of Use

• How easy it is to set up and test the solution
• User Interface - Functional interface should be sufficient to get a pass score.

5. Performance on new/unknown data

• How well does the solution perform against new data?