Abstract: Current technology development in mobile computing and upcoming application stores enable an easy development and distribution of mobile applications. This leads to an increasing number of available applications and to the user’s problem of content discovery. Recommender systems aim at guiding users to relevant items. Currently, recommender systems for mobile applications neglect that the usage of mobile devices is characterized by perpetual changes of a user’s context. In this paper, we give rise to the context-aware recommendation of mobile applications. We explore the design space of recommender systems for mobile applications and describe the different dimensions and techniques for capturing the users, the items, the contexts and the corresponding relevances. For proof of concept we present the prototype of a recommender system that combines the design options in a hitherto unexplored way.

You will find the approach described in the paper and some other ideas that came up and have been reported to us in the last weeks within the next version of appazaar. It will mainly improve our recommender system and fix some bugs of the Android app.

While you’ve been looking for new apps using the lastest version of appazaar, we baked a new version with a lot of improvements and integrated a lot of your feedback as well as our own ideas into it! And now, after a long time of testing, it is finally available in the Market

This is what the new mobile app looks like:

This are our changes in detail:

• “What are you just doing?”: Our recommendations are now tailored to your activity! Just answer the question and tell us something about your current activity to get the right apps.
• Tags for your activity: So far you had to insert free text to describe your activity. Now you can use tags to describe your current context more quickly. And in addition you can now connect appazaar with Facebook! The tags you insert into the widget are directly posted to your status on Facebook
• Blacklist: An app recommendation appears all the time and you don’t like it? Now you’re able to get rid of it by adding apps to a blacklist which can be edited within the settings.
• Scrolling: You are now able to scroll through our recommendations from page to page directly within the widget. We think, this is a good tradeoff between widget size and the number of recommendations. What do you think?
  
• Battery usage (GPS): Moritz did a really great job to improve the performance and battery usage. Nevertheless, for now we can not completely abandon the use of GPS positioning. Specially in areas with bad wireless/wifi coverage or insufficient accuracy appazaar falls back on gps positioning. There are already some more ideas in the pipeline to reduce this further and we’re confident to be able to minimize the usage of GPS further with the upcoming versions!

So just update appazaar or download it via the Market Store and let us know what you think about it

(If you experience some issues after updating, please try a clean reinstall before submitting a bugreport!)

In the meantime we have some great news: We will present appazaar at Pervasive 2010 – The Eighth International Conference on Pervasive Computing. We got our poster “Contextualizing Mobile Applications for Context-aware Recommendation” accepted for presentation and the reviews are very motivating! Here’s a short summary of the papers content:

Abstract: Current recommender systems for mobile applications neglect the very mobility of their users and their perpetually changing contexts. Although applications naturally serve for a special purpose, they cannot be addressed to a particular context of use. Hitherto, context information is not taken into account for the recommendation of mobile applications. In this paper we present work in progress on a platform that gives rise to context-awareness for mobile application recommendation. It records the users’ application usage and relates it to context information that is traced on the users’ devices. Thereby we can determine the context of use of the applications. We describe the conceptual design of a recommender system that exploits this information and suggests only contextual relevant applications to its users. A prototype implementation is presented.

If you are interested in all the techniques and algorithms running behind the software, you can meet us at Pervasive 2010 in Helsinki, Finnland. You’ll also find the paper and the poster on this website after the presentation at the conference. However, if you cannot make it to the conference, please drop us an e-mail for any feedback, question or suggestion.

Oh and by the way, the next release is coming soon, we are just in the phase of stabilizing the new features. And it’ll come with a lot of new features and improvements

Changelog:
*new layout
*less battery usage
*less gps usage
*performance tweaks

Oh and I shouldn’t forget to mention that we’re also planning on releasing the upcoming version in the Android Market Store as well!

Stay tuned