This year, we’ve thrown our hat into the ring for the “Forrester Groundswell Awards”, which recognize social technologies that advance organizational goals. We’ve nominated our semantic tool built to search through Twitter in the B2B category. Our technology helps companies “hear” what customers are really saying. Please read on to get a quick overview of the Twitter website, a look at how it works and some additional information about our involvement in the Twittersphere.

Background on the Expert System Twitter Semantic Search Site
The Twitter semantic search application is designed to quickly and accurately isolate relevant information in hashtags, conversations and individual tweets. Whether Twitter users are discussing a relevant product, service or issue, companies need a way to monitor these conversations to see what the public is saying and make informed business decisions.

Because our Twitter search site leverages our COGITO semantic software, it’s able to interpret tweets with natural language processing and understand their nuances as a human would, all in real-time. To give a few examples:

  • The search application is able to equate the terms “huge traffic jam” and “hge trfc jam.”
  • It recognizes differences in sentiment and cause and effect. A tweet that states “high blood pressure contributes to heart disease” is interpreted and recorded differently than one that states “heart disease contributes to high blood pressure.”

These capabilities give a vast advantage over the keyword search model, which eliminates relevant results that don’t contain a few exact words, and the inefficient manual search, which is costly in terms of time.

For additional information on the Expert System Twitter Semantic Search site, please refer to the press release announcing the site. You can also read this article from Semantic Web, in which I offer some additional insight into the problems that the Twitter website solves as well as practical uses for this technology.

How the Twitter Semantic Search Site Works
In order to analyze the results of a particular Twitter search, the web based service generates a graph to visually represent its findings. For example, below is a graphical representation of the tweets about Disney, which gives the viewer a real-time snapshot of how the company is discussed in the Twittersphere.

Expert System Twitter Semantic Search, Disney Chart
More specifically:

  • The X axis represents time, underscoring the website’s real-time search results. In an effort to keep results as current as possible, each search looks at tweets from the past hour.
  • The Y axis represents the quantity of tweets that mention Disney in a particular context.
  • The dots on the graph represent the range of subtopics that relate to the main search term. For this graph, we can see that Disney was discussed in the context of cinema, TV, tourism, music, commerce, etc.
  • The size of the dots represents the number of users that have covered each subtopic.
  • With this in mind, each dot’s position along the X axis indicates how recent tweets about a particular topic are, and its position on the Y axis demonstrates how strong the association is between the search term and that particular subtopic.

The same graphs can be generated to view analysis of a particular hashtag or for the individual tweets of one influential Twitter user. The site can also generate user maps and topic clouds. To get a better sense of what this would look like, let’s examine the following graph, which represents Lance Armstrong’s individual tweets over the course of one hour.

Expert System Twitter Semantic Search, Armstrong Chart

  • As with the Disney graph, the X axis gives us a timeframe—when during the last hour Armstrong tweeted about specific topics.
  • The Y axis indicates how common certain topics were among Armstrong’s tweets.
    — Specifically, of the topics that Armstrong discussed in the last hour (identified in the legend below the graph), the Y axis indicates that themes like sports, work and finance were the most popular.

For additional insight, please refer to our online demo of the site’s functionality.

COGITO Semantic Software
With the Twitter COGITO Search now in context, let’s backtrack for a moment to get a closer look at the COGITO platform, the software at the core of all our technology solutions.

  • COGITO is a revolutionary semantic indexing, search and analysis tool for the management of strategic internal and external information.
  • Unlike most search technology, it automatically understands the true meaning of words in the same way that humans are taught to read and understand.
    — By understanding the context and relation to other words in a sentence, COGITO ensures the correct meaning of the word is interpreted.
    — For example, the software is able to differentiate between jaguar the car and jaguar the animal.
  • These capabilities allow users to monitor, search and automatically correlate information crucial to their businesses, pulling data from multiple sources including databases, documents, articles, email, Web pages, etc.

Keeping this background in mind, you can also read thorough this article, to take a deeper dive into the relationship between semantics and social media.

For Your Reference
To see the other ways in which we engage in the Twittersphere, we invite you to visit some of our senior staff on Twitter:

J. Brooke Aker, CEO of Expert System USA – http://twitter.com/brookeaker
Luca Scagliarini, VP Strategy at Expert System – http://twitter.com/scagliarini
Rita Joseph, VP Federal Sales at Expert System USA – http://twitter.com/msritajo
Julie Hartigan, CTO Federal Programs at Expert System USA – http://twitter.com/juliehartigan
Bill Porter, Business Development Director, Expert System UK – http://twitter.com/pointer5005
Francesca Spaggiari, PR and Marketing Manager at Expert System – http://twitter.com/fspaggiari


Share On

Menu