Design, query and evaluate information retrieval systems.
Introduction
Many of us have used information retrieval systems without thinking much about what goes into making one. The ones we use now are often so good that we can retrieve relevant results after just one search. But what happens when we don’t get what we are looking for? There have been times when I have not been able to get results, and when this happens there are a few questions I ask myself: Is it possible that what I’m looking for doesn’t exist? Is the database not smart enough to retrieve good results? Am I not searching correctly? In the process of getting my MLIS, I have been both a user and a creator of databases, and I have found that knowing how to design, query, and evaluate these information retrieval systems has helped me to use them more effectively.
I have designed two information retrieval systems so far: one for keeping track of items in my refrigerator and another for cataloging my small but growing art collection. In both of these cases, I had to do some planning before I could get down to designing my databases. I thought about who would be using them (just me and maybe my husband), how I would use them (would I be tracking expiration dates, quantities, insurance values?), and what kinds of information I wanted to include. From there, I did some testing. Since I knew what was in the databases, it was easy for me to tell if my queries missed something and I could make changes to get better recall and precision. Once I finished testing, I could evaluate them and see if they met my needs. I feel as if the evaluation process for these two information retrieval systems is ongoing because my needs might change and what works for me right now might not work in the future, but so far they have been working well.
Evidence
Information Retrieval Systems Questionnaire
My first piece of evidence is a series of twelve questions intended to gauge my understanding of information retrieval systems. The questions are related to all aspects of working with information retrieval systems and such topics as classification, natural language searching, targeted searching vs. browsing, interface design, and the relationships between designing, querying, and evaluating these systems.
My second piece of evidence is a FileMaker Pro database I created to catalog my art collection using the VRA Core 4.0 metadata standard. In planning the database, I thought what I might use it to search for entries and what would be most helpful, so I created controlled vocabularies for some of the fields in order to retain consistency when I make new entries. One of the reasons I chose the VRA Core standard was that I could associate multiple images to a single record. Since some of my art has inscriptions or other markings on the back, it is useful for me to be able to see all associated images of a work when I search for it. This was my first attempt at making this type of database, and I am thinking of ways to improve it. For example, I only have room for one creator in a record, and no way to show if something had one or more artists. I have a feeling that I will constantly be evaluating this database as my art collection grows and changes.
Professional Development
Working with information retrieval systems is something I will continue to do throughout my career. I already have those that I prefer to use, but in order for me to be able to provide the best service to library patrons, I know that I should use all of the ones available and get a feel for which ones will be best suited to which audiences. I can do this by checking the indexing, controlled vocabularies, or the ease of use. Knowing how each information retrieval system works will make it easier for me to make appropriate recommendations.