An Android recommendation system applied a decision science method to a problem people solve badly every day.
Where to eat is a decision with too many inputs. Price matters, but so does distance. A high rating is worth something, until the queue outside makes it worthless. People weigh these factors differently on different days and rarely agree with each other.
A team at Universitas Brawijaya built an Android application to handle that weighing. Their approach rests on the Analytical Hierarchy Process, a method for turning messy multi-criteria decisions into a ranked list.
AHP starts by breaking a decision into criteria. Then it asks a simple question over and over: which of these two matters more, and by how much? Answering hundreds of small comparisons is far easier than assigning absolute scores, and the answers combine into a weight for each criterion.
The method also checks itself. Once the comparisons are in, AHP calculates whether the answers contradict each other, and flags a set of judgements that cannot all be true. That consistency check is what separates it from simply averaging a survey.
The decision happens while standing on a street, not at a desk. Building the system on Google Android puts it where the question is asked, and a smartphone already carries one of the criteria that matters most: location.
The paper sits at the start of a line of work from the same group on mobile culinary recommendation, which later compared AHP against hybrid variants. Matching what a human expert would pick stayed the difficult part throughout.
Aryo Pinandito, Chandrawati Putri Wulandari and Ratih Kartika Dewi (2015). Culinary Recommendation Systems Using Analytical Hierarchy Process on Google Android Platform. International Conference on Engineering and Information Technology (ICEIT).
