FUZZY TEXT ENTRY

Aus DPWiki - The Design Pattern Wiki

Illustration

Google [1] uses fuzzy text search to present users alternative or corrected search phrases in the "Did you mean:" -function

Context

…you want to provide an easy to use TEXT ENTRY function.

As text entry on mobile devices is very fault-prone, you want to implement some mechanisms to avoid or ignore errors.

Problem statement

Entering information on mobile devices is very fault-prone. How can text entry be made more fault-tolerant?

Problem description with forces

Entering text information on mobile devices with poor input means can be very fault prone. Single keys on a PDAs or mobile phones keypad are very close together and with input means like triple tap typing errors often occur.

Even if entering address information causes no faults, sometimes users just do not know how to spell a street name correctly for example.

Although the first fault source can be identified and corrected, this requires additional thinking and action of the user whereas the second source leaves the user alone with the problem.

For this reason a mechanism that automatically recognizes faults and presents users corrected search results is needed.

Examples

The idea of fuzzy logic was first introduced by Lotfi A. Zadeh [2] in the 1960's as a means to model the uncertainty of natural language. Fuzzy logic is a superset of conventional (Boolean) logic that has been extended to handle the concept of partial truth - truth values between "completely true" and "completely false".

As shown in the opening picture, Google uses fuzzy text search to present alternative or corrected search phrases to users in the "Did you mean:"-function.

Some navigation systems also use fuzzy text entry for entering address information. SMARTPATH City [3] for example shows a list of possible results while the user is still entering address data.

The list shows results that contain either the - so far - entered search term or that are similar to the term. In the example given the application ignores the wrong spelled "noble" and displays street that include the searched street "Nobel-" and also alternative spellings like "Nobile-" or "Nebel-".

Solution

Therefore:

Make your text entry function use fuzzy text search.

With this fuzzy text search input faults like typos, equivocators or other possible faults can be recognized and eliminated. So users do not need to perform additional actions to correct their mistakes as the system takes this work from users automatically.

Either present the corrected words with ignored faults immediately in a selection list or let users choose from a list with correct addresses after search.

Diagram

References

This is a basic pattern and has no further references in this pattern language.

Literature and Links

1. Screenshot taken from www.google.com in June 2003
2. Zadeh, L. A. (1965): Fuzzy sets, Inf. Control 8, 338-353.
3. SMARTPATH City, Version 1.10 (161); http://www.navman-mobile.com/html/city-pc.htm