User Profiling and Social Network Analysis
Our profiling engine can learn demographics and niche interests of your customers or visitors from text or item-based data.
- Text-based inputs may include web-sites that a user has visited, or comments they have made.
- Item-based inputs may include the list of items that a user has viewed or purchased.
- Profiling can take into account time and location.
The user-profiler can be setup to classify users into demographic boxes (e.g. by gender, approximate age, earning bracket) and can provide detailed, machine-readable profiles of each user’s interests and habits. The more information that is available on the user, the more detailed and reliable the profile becomes.
The user-profiling engine is not only applicable to retail scenarios. It is relevant to any company wishing to:-
- Increase revenues through direct advertising.
- Extract valuable market intelligence.
- Spot trends among their customer or user-base.
ThinkTank’s core technology is a specialist software toolkit, including an extensive Natural Language Processing (NLP) library coupled with state-of-the-art statistical learning and inference algorithms. Our framework allows us to rapidly develop, test, deploy and maintain sophisticated systems. It is highly scalable and works with many languages besides English.
Social network analysis
ThinkTank has developed a set of social network analysis tools, building upon cutting-edge research in complex network theory.
The tools can be used to infer the strength of social relationships between individuals, automatically discover communities and determine the amount of influence or “centrality” an individual has. One possible application of this technology to on-line environments is the ability to identify “hub” users – the drivers of the community who are ideal targets for marketing offers. Such techniques could be applied to any service with social aspects.
Whitepapers
Please contact us at to obtain access to our whitepaper “Seizing the Long Tail”.
