WEBSOM - Instructions

What is WEBSOM?

WEBSOM is a method for automatically organizing collections of text documents and for preparing visual maps of them to facilitate the mining and retrieval of information. The documents are in the points, or "pigeon-holes", of the map, and their contents can be browsed by clicking the points visible on the lowest level of the map display. You can use a full text search to find an interesting starting point for browsing.

Uses

Browsing

Click on the map image with the mouse. A zoomed view of the part of the map will appear. By clicking on a white point you can investigate the contents of the individual map unit. The arrows let you move to adjacent units, where you will probably find similar texts. You can read the texts by clicking on the headings.

The map image contains labels which are example of the core vocabulary of the area in question. The labels give a general idea of the topics in the document collection.

The colouring of the various areas of the map represents the density of documents on that area. Light areas contain more documents.

Search (Only available in some demos)

You can write a description of what you wish to seek in the search field. The description can be a few keywords or a freely worded sentence. You may choose between two search modes, Keyword search or Document search. In Document search the program encodes the search text as a document, and finds the area of the map with most similar documents. In Keyword search mode the words are combined either with AND or OR (select either one), and the best-matching locations on the map are returned.

After you've pressed "Go!" the program will seek the area of the document map with the best match for your text and will mark them with a circles. By clicking on the circle you can browse the results of the search. The size of each circle represents the goodness of the match.

Try for example coffee machine. The search facility is particularly useful when looking for an interesting starting point for browsing.

If none of the words in your search text are found in the material used for the map, or if the words are very common or very rare, the system will report that no search can be carried out. Try a different kind of description.