KCSL

Classification – Categorization, Taxonomies, and Meta Tags

Myth

Stemming from the mindset created by the card catalogues of old, it is thought that by classifying documents into particular categories, the task of finding documents becomes simpler, and more effective, than would finding documents through a search box alone. The categories, which are often based on Meta Tags, are designed to be navigated by users clicking through numerous category levels to find the documents they need. Categories are also used to narrow the scope of user searches; the idea being that the search system will perform better by only having to analyze a particular category or categories for a specific search, as opposed to all categories for every user search.

Fact

Two people seldom, if ever, categorize the same document the same way, thus users frequently go astray somewhere along the categorized path to the document they are seeking. Also, by narrowing the scope of a search, users run the serious risk of missing a relevant document that may have been placed in a category they aren't searching.

With a system such as UniFind™ Technology, a simple Natural Language query can always find the most relevant documents from amongst all the documents, and hence predefined categories and taxonomies are obsolete. With UniFind there's no need to spend enormous time and effort creating and maintaining categories, nor to complicate the search process with them. In a sense, UniFind works so well that it automatically performs dynamic categorization every time a question is asked.

Back to Technology : UniFind Search : Unique Approach