Maja+-+Annotated+Bibliography

Process flow
 * organizational process that the tech streamlines (facilitates inventory)
 * it enables cross selling
 * supplements marketing
 * helps increase product awareness (as part of marketing strategy)




 * Wei, Yan ZHENG** **; Moreau, Luc ; Jennings & Nicholas R "A Market-Based Approach to Recommend-er Systems //." ACM Transactions on Information Systems.// 23.3 (** **2005)****: 227-266.** **Print**

This article elaborates on the idea that the World Wide Web imposes an information overload problems on the side of knowledge workers, that seek the best way possible to reach all customers, as well as for customers who seek for the best information out there, with the preference of the least time and effort invested. Recommend-er systems, where implemented make recommendations on behalf of the user from a variety of sources without any pre-knowledge needed.It has been established, that so far their value lies in the ability to aggregate information from past uses and present it to the right user .The article proposes integration of the so far prevalent types of recommend-er systems: the content-based and the collaborative filtering recommend-er systems of information into a market system. The goal is to overcome some of the drawbacks of the individual systems as well further improve them with additions, e.i such as demographic filtering of information (to further refine the results shown to users).


 * Weng,****Sung-Shun****; Lin,** **Binshan** **and Chen**,**Wen-Tien "** **Using contextual information and multidimensional approach for recommendation**." //**Expert Systems with Applications.**// **36.2.1 (2009): 1268–1279. Print**

The above cited article elaborates the concept of multidimensional recommend-er system. It looks at the possibility for integration of multiple dimensions of data such as profiling, aggregation, and multifaceted capabilities. The authors puts emphasis on the power of the online analytical processing (OLAP) and their potentials in resolving many problems that may arise as a result of contradictions among the variety of hierarchical ratings. The author makes visual presentations of multidimensional cubicles of information. By implementing this, the occurrence of difference in ratings and the problems they impose in recommend-er systems (due to the fact some purchase decisions of mainly services are not solely based on preferences among users, but may take into account the actual environmental factors in the decision making process such as restaurants and travel) can be resolved. In addition, a movie recommend-er prototype system is being presented, with all the details about the structure of experiment system and explanation; setting of system parameters; process of the experiment and experiment result and analysis.  **Weng,Sung-Shun ; Liu, Mei-Ju "Feature-based recommendations for one-to-one marketing." //Expert Systems with Applications// 26.4 (2004):493-508 Print.**

Currently, many web sites are selling products that are aimed at very specific, essentially very small segments of the market ("the long tail"), and thus the recommend-er system (which is based on the premise that many other customers have seen the same or similar products) is not very much suitable for them.The article elaborates on the need for companies to employ practices that can make new and products that are not a frequent buy (such as furniture) more visible to customers through recommendation (if they want to stay profitable).As the article proposes, building personalized interest profiles for customers from purchase history based on the features of the products purchased can be the basis of a more successful recommendation systems, resulting in increased accuracy in recommendation.


 * Lee, Shao-Lun "Commodity recommendations of retail business based on decision tree induction."** **//Expert Systems with Applications 37.5 (2010):////3685-3//694 Print.**

This article presents the concept of collaborative recommend-er systems in their role in commodity recommendations. It gives a suggestion as to how collaborative filtering can be further improved. It proposes data mining and analyzing retail business data as a methods to make data more relevant and customized to customers. It presents some of the weak sides of RFM scores, after going trough the drawbacks of some other commonly used methods to generate correlations among customers. In addition, it proposes NRS, a new proposed method which overcomes the negative sides of the earlier. As final, it proposes the use of a decision tree to explore all possible associations among commodities.


 * Porcel,C.; Moreno,** **J.M.& Herrera-Viedma****,****E.** **"A multi-disciplinar recommend-er system to advice research resources in University Digital Libraries.**" **//Expert Systems with Applications// 36 (2009): 12520–12528** **Print**.

This article is about the implementation of a recommend-er systems in the digital libraries systems. Its proposed goal is to embellish already sophisticated digital library systems with recommendations options in order to cater to diverse communities interested in also diverse research aspects. The system planned to be put in place is called "fuzzy linguistic recommend-er system" and aims at recommending users articles that may be of their interest. In that instance, new users can be offered newly added articles as well as given the opportunity to connect to other scholars based on common interest.The underlying idea of the article is to keep the advantages of collaborative, content, demographic as well as knowledge based filtering of information and form a superior hybrid system.