In this article, authors Vairale and Shukla concern themselves with so called „recommendation frameworks“. These are structures that gather large quantities of information and then proceed to recommend specific items aimed at groups or individuals. Those groups or individuals are assigned to products based on their personalized profiles which include info about specific inclinations and preferences, likes, histories and so on. They can work in several different ways. The first way is so called „colaborative filtering“, which makes proposals based on user ratings of items and so generated group recommendations. Another way is „content based systems“ which compare information about the product directly to the info about the client. Third way seems to be the most effective one and it relies on the perpetual loop of considering client data, recommending item, looking for feedback, adjusting profile and then looking for a new feedback. The last method of frameworks operation is by a combination of previously mentioned methods.
Frameworks like this are mostly used for marketing purposes. In this paper however, authors tryed to identify projects that used these frameworks in order to enhance users health. There are two big cathegories in thi field: diet and exercise recommendation frameworks.
This study identifies several projects that create personal profile of each user, do extensive background checks for issues in medical history, food preferences, lifestyle choises and so on and try to recommend various objects such as books, films, food items, menus and workout routines to help improve subjects overall wellbeying. Algorythmic approach applyed to this issue can potentialy be of great help to those of us who need to fight specific disseases or are sensitive to certain food items. It can also help alleviate the pressure of living in high information density environment. Information paralysis itself often stands as the main reason why we are not able to change our lifestyle and limiting information into a managable stream can help us deal with this issue.
In my oppinion, this field is very interesting and has a great potential with further developement of technologies. For example I have not discovered anywhere in the article mention of wearable technologies for tracking different biofactors and using them to adjust recommendations. Today this sort of technology is very limited due to self-reportig nature of data harvested from users. These tend to be biased or inacurate and tend to demand great willpower on the part of participants.
Sticking to the recommended guidlines is another issue that cannot be adressed by these frameworks. The article sadly does not go deep into specific projects. It just mentions several of them and describes them in a few sentences. That is a great shame, because I initially chose this specific chapter in hopes of finding good tool for myself. On the other hand it serves well as an introduction to this area and presenting the concept of recommendation frameworks.