Why Serendipity, Personalization, Curation and Prediction Analysis Matters

December 16th, 2010 by Mattias Tyrberg

Serendipity, personalization and curation are all buzzwords that we see more often. Saplo is building a text analysis API and we thought that methods for personalization and context analysis would be useful for our developers. This blog post will give you a few ideas on what contextual analysis can be used for.

How is this valuable?
Today more than ever we spend huge amounts of time analyzing text only to decide things like:
- Is this information relavant to me?
- Who should I pass this information along to?
- What is this information mainly about?

We think these kind of decisions can be made by computers, letting us focus on what really matters- reading and learning from content we find useful. Here is some suggestions on cases when contextual analysis could be of great use:

  • When you as an insurance company receives an insurance case by mail you would want that mail (with unstructured text) to be passed to the right derpartment and person instantly.
  • When you as a news site publisher publish a new article you would like that article to automatically be placed under the right section and be labeled with relevant topics.
  • When you and your support team receives support requests you want to be able to find similar questions and answers that might be useful in this specific case.
  • When you want to have contextual ads and dont want to use keywords. Semantic technologies can be used so that you only need to choose a number of articles that you would like your ads to be displayed next to and Saplo Context would find other articles that is similar to the ones you’ve chosen and display your ads next to those as well.
  • When you would like to find an interesting article. Like when Amazon is recommending items to buy based on your interests but with text articles instead. This can be done by using Saplo Context to create a personal context defining your interests. Then you will get personal recommendations based on what you find interesting. The ideas of where this could be useful includes sites such as blogs, news sites, RSS aggregators, search engines etc.

So what is Saplo Context?
A context is created from one or many texts. In example, three different sport articles can together build a sport context.
After a context have been created it is possible to compare other texts and articles to that context. You will then get a value of how similar the text is to the context. If you compare a new sport article to your defined sport context you will get a high result value indicating that the article is similar to the context, but if you instead compare an economy article to the sport context you will get a lower result value indicating that this text is not as similar to the context as the sport article was.

Do you have other ideas of how to use it? Feel free to add a comment, send us an email or tweet us @Saplo.

If you would like to try our contextual analysis, please sign up for Saplo API.

Note: Saplo now support contextual analysis for English and Swedish. If you are interested in analyzing other languages please contact us. Since Saplo is using a language independent technology we can add more languages with your help.