INLS490-154W: Information Retrieval Systems Design and Implementation

This is the podcast for the thirteenth class, in which we look at a couple of ways to organize and present information to the user. We see how a term-cloud interface can be created, allowing the user to get a quick glance at the underlying collection. We also talk about a number of clustering algorithms and see how they can be implemented with Lemur.

This is the podcast for the twelfth class, in which we see how REST requests can be made through the web, and the responses in XML can be parsed. This allows us to start connecting with Web 2.0 sources that provide the functionality of meshing different sources by open data exchange.

This is the podcast for the eleventh class, in which we see traditional and non-traditional methods of collecting data off the web. Traditional way is demonstrated using web crawling using wget, and non-traditional way is instantiated with YouTube harvesting.

This is the podcast for the tenth class, in which we connect the back-end for search that we have been working with to a web-based front-end. This is done using Indri, a new search engine component for Lemur. We also explore some details of AJAX and see how we could use it to enhance our user interface for search.

This is the podcast for the ninth class, in which we continue looking at evaluation. We talk about more measures to evaluate a query and a system. We also look at comparing two rank lists.

This is the podcast for the eighth class, in which we start looking at one of the core components of IR - evaluation. We begin our discussion by revisiting recall and precision, and then continue exploring R-precision, AP, and MAP. We see how these can be measured manually and then using TREC supplied tools.

This is the podcast for the seventh class, in which we look at how structured queries with term weights can be executed using Lemur. We use this to provide terminologic feedback to the user, and incorporate the relevance feedback that the user provides into retrieval process.

This is the podcast for the sixth class, in which we continue exploring models for retrieving information. So far we have seen vector space, boolean, and language models. Now we talk about probabilistic and relevance models.

This is the podcast for the fifth class, in which we dive into models for retrieving information. We had already looked at the vector space model. Now we talk about boolean retrieval and language models for retrieval.

This is the podcast for the fourth class, in which we look at how queries can be processed using Lemur Toolkit. We then match the processed queries with collection index to retrieve a rank list.