Current search methods rely heavily upon the user’s ability to provide the correct keywords to convey the concept they seek. This leads to several problems:

• Words, particularly acronyms, have multiple meanings leading to many irrelevant search results. This is compounded by the fact that almost 50% of all searches use less than three words. For example, does “EDC” mean Electronic Data Capture or Economic Development of Canada or Education Department of Colorado any other number of possible designations? Does a person interested in searching for “cancer” (as a medical term) want to have their search results populated by horoscope listings (or vice versa). Is the person searching for “ Sweet Home Alabama” looking for the Reese Witherspoon movie or the Lynyrd Skynyrd song?
• Large records, such as lengthy books and encyclopedias, might contain the specific words themselves but be largely irrelevant to the content sought or contain only a small section of relevant content in a much larger document.
• Top down or “tree branch” classification systems will often miss relevant secondary or tertiary topics that are contained within a document but are different from the primary topic under which the document was classified.
• Documents that contain synonyms of the specific chosen keywords may not be found. Suppose the key words “small lake” were entered but valuable information was missed because a very relevant document containing the term “pond” was not found.
• Documents written in foreign languages will be largely inaccessible.