Statement of Competency E

Design, query and evaluate information retrieval systems.

     Information retrieval (IR) systems, such as a library OPAC, have replaced the antiquated card catalog in many libraries. The OPAC is only one example of the myriad of an IR system used in the library profession today. These systems focus on delivering information to a user’s query by returning a set of results. These results must maintain a balance of recall and precision acceptable to a user’s needs. Today’s librarian must be versed on Mooers’ Law concerning information retrieval; keeping in mind creating an IR system means understanding the target audiences. This understanding will help build a system that will retrieve appropriate information on demand. The information seeker may go elsewhere if their queries are not met in the fashion deemed acceptable. I have found a working understanding of information retrieval system design, query and evaluation has opened many potential job opportunities for me as a library professional. As a Business Analyst for an IT department, I found application of my current skills to library science a perfect union when studying IR systems.

Design

     Information system design begins with a focus on the user. Designing the system around a specific user model ensures the information seeking needs will be met. User centered IR system design should concentrate on who will be using the system, how it will be used and what is the goal to be achieved by the system. After the user model is resolved, the data structure of the IR system database is determined and surrogates made. Description of items in the systems should follow conventions the potential users are likely to search.  

     In the submitted evidence, Information Retrieval Database Design, from LIBR 202 (Information Retrieval) my group explored the user model and proper design needed to deliver information on a specific collection to our users. The database was designed using Inmagic DB/TextWorks software. Our first steps were to brainstorm on our potential users and the reasons behind their need for the information. The “data structure” and “records” section of the assignment illustrates our surrogates. The data fields and indexing rules were based on the probable search strategies our user model would initiate. I was able to exchange records with another group, give them a set of instructions and illicit their feedback on the usability of the design. This assignment helped me develop an understanding of proper design of IR system database structure.

Query

     An understanding of the queries a user is likely to use is necessary when designing an IR system. This will determine the sophistication of indexing terms applied. For instance, an IR system based on a marine biology professional user model might want to index “coral” on a level including species type, time periods first appeared, and other highly detailed search fields that would not be included in a more generic database. The “Information Retrieval Database Design” evidence illustrates the level of indexing terms used for our specific user model.

     The ability to query an IR system does not only include understanding of user models in terms of IR design. The submitted evidence, SQL Query, from LIBR 242 (Database Management) illustrates queries to a database management system (DBMS), which is at the core of an IR system. SQL is used when querying and controlling data in a databases. The ability to query the database of IR systems via SQL is helpful when attempting to create reports or retrieve specific data from a library’s DBMS in the IR system. The assignment shows the SQL queries I would use in response to querying a database of published books. Knowledge of this level of system query has provided me with a valuable skill opening my employment options to include Data Management and Information Architecture fields of library science.

Evaluate

     Evaluation of IR systems helps determine the success rate for which the user queries are met. The effectiveness of the system is determined in its ability to retrieve relevant results while withholding non-relevant information. The response time of the system to return a query is also valuable to determine. Slower systems and irrelevant results will result in user dissatisfaction. Recall and precision are two measurements used in determining the effectiveness of an IR system. Chowdhury (2004) explains users prefer a level of recall of 60% while high precision results in saved time and effort by the user (p. 250).

     The submitted evidence, Vocabulary Design – Evaluation, from LIBR 202 (Information Retrieval) shows the evaluation portion of a group project for which records were created for journal articles. Once again, a user model was determined and possible searches defined. After creating the records, searches were conducted using specified fields. Finally, recall and precision are calculated for each field. This assignment helped in understanding how well the records were built and what could have been done better to bring the user more accurate results.  

References

Chowdhury, G. G. (2004). Introduction to modern information retrieval (2nd ed.). London:Facet.