Solutions

Twitter Language Tools

These tools are the results of our R&D effort in building solutions for the Twitter users and marketers communities. The tool set contains a:

• Twitter Retrieval Library. This tool helps in finding relevant Twitter messages or users according to target keywords. Extracted information will be stored in a local database and can be used for further analysis.
• Twitter Text Corpus. This tool is based on the processing of large sets of messages extracted from Twitter’s real-time message stream. This tool is delivered in the form of a database that contains the most used expressions on Twitter. The tool also provides the frequency of each expression in all Twitter messages as well as in a set of predefined subject areas. This database is clean of general and Twitter-specific stopwords.
• Corpus Management Library. This tool is used to maintain the corpus with new expressions. This tool aims at solving the problem of having a contemporary corpus by updating it with new messages as they are created by Twitter users.
• Twitter Clustering Library. This tool will use cluster analysis methods to group similar Twitter messages or users together. The techniques used here are based on unsupervised learning algorithms and aim at finding patterns from data extracted from Twitter.
• Message Categorization Library. This tool will help in classifying Twitter messages by subjects such as Business, Technology, Sports, Celebrities, etc.
• User profiling Library. This tool will classify Twitter users based on their timeline (messages, followings).
• Message Contextualization Library. This tool helps in connecting Twitter messages with external sources of information. The library will analyze Twitter messages and link them to webpages from as news websites and blogs.

Patent Search Tools

This tool set is designed to provide patent agents with a simplified and more productive search experience when conducting research on prior art. The following features:

• Standard Search: users can search patents with standard options such as keywords, inventors, cities and assignees.
• Patent Clustering: search results can be clustered using different classes of attributes such as patents texts, citation networks and classes. For each cluster, an automatically label is assigned and major assignees and inventors are identified.
• Citation Networks: this feature will identify patents that are central and that have a high technological impact.
• Relevance Feedback: users can specify which patents are relevant after running an initial search. The software will analyze the specified patents’ texts and extract significant terms that will be suggested to the user. The user can choose between suggested term and run a new expanded search.

Contact us for more information about our solutions.