Changing Web Behaviour
Problem being addressed
Over time user behaviour on the internet changes depending on interests, age and other factors. Can the evolution of a user's web interactions be modelled and understood?
User behaviour on the web tends to vary over time. This behaviour includes web queries, favourite sites, and social media activity. In this paper, a time varying model and approach is created for the purposes of understanding the dynamics of user behaviour on the web. The key ingredient is a learning model that can be used to construct models of user's behaviours (url and search query behaviour) based on current and historical activities. Aspects of the model include predictive forecasting for trends, disruptions, and periodicity. Further, a novel model selection algorithm is developed in the paper for choosing the best model out of a list of models, when confronted by different patterns of user behaviour. This is shown to be better than using a static one-size-fits-all approach.
Advantages of this solution
This is an interesting piece of research that looks at web dynamics over time. It has applications in the internet sector in terms of forecasting and predicting consumer behaviour.
Solution originally applied in these industries
Possible New Application of the Work
The media has a massive influence on society and this influence can be understood in terms of what people look for information about. One application of this research could be to connect internet and media related queries, and use this to gauge people's sentiments over time. This type of analysis could help break through echo chamber effects that are a danger to democratic society.
In sports player careers tend to be time varying storylines, and different players tend to take on different roles as they get more experienced and age. This is true, for example, in soccer and in cricket. Perhaps, sports analytics could be used with time varying analysis to predict and shape the careers of younger players so that they end up having longer, less injury prone careers.
Travel and Tourism Industry
This research could be applied in the online travel space to understand the relationships between user questions on search engines, and their ultimate end point in terms of travelling destination. Since holidays tend to be "rare" regular events that follow time varying patterns, this research could provide interesting insights into holiday choice behaviour.
Source DOI: #############