Automated Recommendation Rule Acquisition for Two- Way Interaction-based Social Network Web Sites

TL;DRAbstract

A problem with social network web sites for activities such asdating or finding new friends is that often there is little positiveresponse from those contacted. In this research we investigatedhistorical data from a large commercial social network site toestablish which subgroups of people were most likely to respondto a particular individual. Our two-way interaction modeldeveloped a table for each attribute to determine which pair ofvalues for sender and recipient gave the best response rate. Fromall the attributes the user profile of a likely responder was created,but then less significant attributes were removed. With this simpletechnique we were able to demonstrate that where users hadcontacted people the system would have recommended, thesuccess rate was 29.4% compared to a baseline success rate of16.6%. This represents a very considerable increase in thelikelihood of getting a favourable response. We are now planninga study that provides prospective recommendations to actualusers

Chat with Paper

AI Agents for this Paper