The amount of digital information available in the Internet and various Intranetsoften causes information-overload, signicantly increasing the amountof time and cognitive resources needed to acquire relevant and accurate information.When searching for information to address complex problems,users spend signicant amount of time clicking through search-query resultsand reformulating the search if the results are not satisfactory. Thisis a tedious and challenging task and has a negative impact on the globaleconomy.Personalisation and contextualisation techniques intend to address the abovementioned problem. Such techniques help to derive additional informationfrom the history of the past user interactions (user prole) or the currentcontext of interaction. This information is used to rene the search resultsin order to narrow down the scope of search queries for better results.One of the key requirements for the development of personalised or contextualisedsearch utilities is consistency of human behaviour. It is onlypossible to predict user future preferences and actions if they are correlatedwith past behaviour. This fact is frequently ignored in the current IR literatureeven though empirical evidence clearly illustrates that humans arevery inconsistent when interacting with information. This leads to very lowpredictive validity of existing contextualisation/personalisation IR implementations.This thesis hypothesises that HIB could be consistent under certain contextualand task specic conditions. The thesis claims that a large proportionof our daily information activities are highly consistent and also meet thedenition of habitual behaviours. In other words, even though empiricalevidence clearly illustrates that on average human information behaviour isvery complex and dependent on thousands of factors, there will exist a smallgroup of daily activities which are highly consistent and can be supportedeectively by modern IR utilities. As a consequence the development ofhighly eective personalised or contextualised solutions are feasible.In order to prove the above mentioned hypothesis User Study 1 (diary study)was carried out. User Study 1 revealed that a signicant proportion of ourdaily information interaction is indeed consistent (49%) and a signicantproportion of this can be classied as habitual (41.9%). User Study 1 alsoconrmed that the behaviour of participants is consistent only when thesame tasks are carried out in the same context and under the same emotionaluser state. Finally User Study 1 conrmed that the behavioural consistenciesare highly individual and aected by a number of external factors.However there exists no HIB model or research methodology which can helpto identify the external factors systematically and take advantage of themduring IR system design. Therefore this research proposes an "IntegratedFramework for HIB in-situ" which intends to ll the above mentioned gapin knowledge.The framework was designed to support the development of Information Retrievalutilities based on consistent behaviour of clearly dened user groupsand their problems. In order to illustrate its applicability a single usergroup was selected. The proposed framework was successfully applied tothe problem of work-related activities of software engineers. The frameworkallowed for identication of a more specic user group of software developersand narrowed down the investigated task to code development/debugging.In the next step the framework allowed for shortlisting a number of behaviourswhich had a signicant potential for consistency. The consistencyof the shortlisted behaviours and their correlation to relevance was veri-ed through User Study 2 (questionnaire). The key shortlisted behaviourswere further analysed through User Study 3 (fully automated, long lastingethnographic study) which allowed for the identication of factors that havea key impact on behavioural variance. The analysis revealed a number ofconsistent behaviours (implicit-feedback indicators) that can be used forprediction of document relevance. Importantly User Studies 1 and 3 validatedthe research hypothesis on a specic case study of software engineers.However the proposed framework is based on very basic cognitive mechanismsresponsible for the human decision-making process and as a consequenceis highly generalizable to other user and problem groups.
展开▼