Wednesday, November 12, 2014

NTNU-ere developer tool that knows what you like – Universitetsavisa

– If we want to create a local newspaper for Trondheim, we can only highlight Trondheim on the map, so. And voila, we get up all the news that is written about Trondheim in the last day.

With a map and a marker that could be moved in both geography and size, showed postdoctoral Jon Espen Ingvaldsen how the NTNU-developed technology can be used to retrieve news written about different places and different themes using a quick search.

Ingvaldsen working with big data analysis and personalization of news, it is to customize our offer to each reader, so you always get the news that is relevant to what you are interested in.

– Technology with commercial interests

Ingvaldsen lecture was held under Nxt Media Conference Scandic Hotel Lerkendal Wednesday. Nxt Media Conference is a collaborative project including between NTNU and Polaris Media, which owns, among others Adresseavisen. The conference focuses on media relevant technology, and several startup companies from NTNU stood on the stand at the conference.

In the same speech also showed scientist Professor Jon Atle Gulla how the Smart Media project at NTNU through both linguistic formulas and semantics is to index news, give them a grade, and then connect them to the readers who fits this character.

The information collected both at the user’s history is analyzed through logs online, and that you put together users in groups. The latter means that if a user, for example, read the same news as another user, you may find that this user is also interested in what the other users have read before.

The research is a continuation of recommendation systems major players already practicing today.

– We know that on Netflix are two of the three films and series seen as a result of the recommendation, 38 percent of news items on Google News is read after being recommended to the reader, and 35 percent of sales on Amazon.com occurs because the products are recommended for customers. So this is a technology with great commercial interest, says Ginny.



Recommend News

The special work done at NTNU, is that working in recommending news for different browsers mobile. Ginny says that there are several problems with this type of recommendation, than it is to recommend movies and books to customers.

– News has several characteristics, including that they produced all the time, and that they have a novelty value that is valid for a specific period of time, they quickly become irrelevant. This stand clear of books and movies, which are objects that are more tidfaste, says Ginny.

The technology makes when you enter the mobile to read the news, you can get a number of articles tailored to your interests, preferences and where you are located in the country.

On its website writes Smart Media that they are about to create a separate open database for Norwegian news, so that others may develop technologies the same direction.

LikeTweet

No comments:

Post a Comment