design, query, and evaluate information retrieval systems.
Librarians are the caretakers and guides to information for the communities they serve. Information retrieval (IR) systems, in all their forms, are the tools librarians use (and offer for use) for information seekers. As an information guide, the librarian should have knowledge of IR system design, a good understanding of efficient querying strategies, and the ability to evaluate IR systems—for recommendation to users with information needs, for relative usefulness in queries conducted for information seekers, or for value provided to the institution.
Before computer systems, librarians were tasked with creating a system to organize their collection so desired information could be easily located. The solution was standardized systems of classification, such as the Library of Congress Classification and Dewey Decimal Classification, as well as catalogs to locate a book in ways other than knowing its specific location. Computers made these catalogs more flexible and useful, and networking exponentially expanded possible sources of information. Today’s IR systems are at times searching through billions of records to find the user’s desired information. Formal IR systems of particular use to the library, such as LexisNexis or Factiva, are powerful tools that require some [know-how] for optimal usefulness. The ubiquity of the Internet also provides more informal sources of IR, such as search engines or web portals for an organization. These types are no less important for an information professional to understand and apply.
Design
Having to create an information retrieval system from scratch is an unlikely scenario. Even the informal IR systems, such as a library’s web portal page, are typically created from templates that can typically be made to serve any number of uses. However, studying what goes into the design of the IR system leads to a valuable understanding of IR system components. The following are a few of the more notable concepts which impact IR system design.
Controlled vocabulary. Establishing and assigning controlled vocabulary in an information retrieval system is an organizational method which serves to classify items within the system. Controlled vocabulary serves not only to formalize the terms used to classify within the IR system but also to advise the user which terms are preferred—tennis shoe over sneaker, for instance. Considerations when creating controlled vocabulary terms concern minimizing the ambiguity between terms. Significant examples, such as the Library of Congress Subject Headings, will list similar terms (variants) which advise to use the preferred term instead..
Hierarchy. Also present in the Library of Congress Subject Headings, subject hierarchy is a method of term arrangement which creates a relationship between search terms which proceeds along greater or lesser specificity. Hierarchical subjects are useful to searchers looking for related subject matter, or to explore an interest more exhaustively.
Tagging. Becoming more widespread with the flourishing of social networks, tagging is a user-generated classification system with which the user assigns their own subjects to a particular item. In an IR system with many users active in tagging, a high number of tags can indicate the veracity of a particular term used. Tagging systems can be entirely user-defined or can be selected from a pre-established list. Popular both as a form of engagement with a user base and also due to savings realized from not having to assign values oneself, tagging can be a wise alternative to a more traditional classification system in the right application.
Query
The most relevant day-to-day skill pertaining to information retrieval systems is querying, or searching. Now a familiar concept in the age of Google, querying is the design of a request to retrieve the information desired. IR systems with rich classification features will often offer advanced search features to narrow results to a specific journal, time period, or other specification. Additional search functions such as wildcards, which are symbols representing a range of results, or proximity indicators, which find a term a certain distance away from another term, may be available. Having an understanding of what can be used to form a search can assist in further directing that search to find the desired information.
As technology progresses, the increased capacity of storage and its attendant decrease in cost, as well as increased processing power, have made it more feasible to offer full text search of resources. Full text search can lead to more relevant results than could be found otherwise, but can also provide irrelevant results if words are included in a document out of the context assumed. The best strategy for finding relevant full text search results is usually a combination of terms to be found within the full text as well as searching keywords or the abstract.
In some settings, most typically a reference desk scenario, librarians will conduct a search for an information seeker based on the information the seeker provides. This can sometimes be challenging because of the ambiguity of desired information (sometimes they don’t know what they want to know). Having experience with one’s own searches in the IR systems can help. An essential skill for creating queries of others is being able to identify key words or phrases in the request and apply them to the search parameters.
Evaluate
The decision to use or not use a resource for information retrieval involves informal, ongoing evaluation. Evaluating whether one’s organization will continue with an IR system, or perhaps will purchase subscriptions with a new system, is a periodic exercise conducted in any effective organization, but can require a more formal evaluation process. In any case, it’s a good idea to evaluate the tools one is using, and the ability to understand principles of design and efficient search will assist in IR system evaluation. Beyond evaluating the design of the IR system and its search abilities, it is important to evaluate what is offered in the context of the community needs. Perhaps even more than content offered, the usability of the interface should be evaluated, as an IR system rich with resources but a difficult-to-understand search syntax will not be well-used if “good enough” can be found with another tool.
Evidence
LIBR 202-11: Information Retrieval—Homework 2A,2B, & 2C
LIBR 202 introduced IR system design with subjects less familiar to library students, i.e., non-books. I chose t-shirts as the subject for my system. This three-part homework assignment consisted of design and presentation of the system along with a statement of purpose, testing and critiquing another student’s system design, and reflection on the critique and project as a whole. When taken together, I believe these assignments show my adequacy at understanding issues related to design, as well as proficiency in testing a system and providing relevant feedback.
LIBR 202-11—Homework 2B note: assignment was conducted in discussion board, the system I am evaluating (another student’s 2A) is below.
LIBR 244-11: Online Searching—Exercise 1
When I told my wife I had a class called Online Searching, she asked “you have a whole class on how to Google?” Even understanding there was more than that, I felt like a full class dedicated to the subject would struggle to fill time. This was not the case, as we conducted complex searches throughout the semester in three separate IR systems, noting the differences and similarities as we went. This exercise is a good example of the thrust of the class, a series of five search problems which I would find relevant documents for in ProQuest Dialog and “show my work” with my query progression, refining as I went for more relevant results.
LIBR 244-11: Online Searching—Exercise 1
Conclusion
In the context of providing relevant documents for information seekers, information retrieval systems are the most important tools available for librarians. I believe that the classes from my work examples as well as others have given me a firm foundation for interaction with IR systems. This knowledge will serve to facilitate searches conducted for others, and will also allow me to instruct others how to better use the IR systems available. In regards to evaluation, I believe I can provide appropriate feedback on current systems as well as systems which will potentially be implemented.
Life is information, information is life. 