Research On Relevance Assessment Of Health Information Search Behavior From The Perspective Of Task Classification

The Internet has increasingly become a leading force in innovation-driven development,profoundly changing people's production and life,vigorously promoting social development,and having a profound impact on the development of international politics,economy,culture,and society.As of December 2018,the number of Internet users in China reached 829 million,with a penetration rate of 59.6%.The “last mile” of network infrastructure was gradually opened up,and the “digital divide” was speeding up.With the deepening of the informationization of medical services in the Internet era,the proportion of users who first conduct online information search in the face of health problems is getting higher and higher.At present,the public's attention to medical and health information is increasing,the demand for medical health information is growing,but the overall level of public health information literacy is low,and the quality of various health information websites needs to be improved.This study aims to explore the rules of health information search behavior of college students' online users in different task situations.The main research consists of two parts.The first part is to study whether different users will tend to choose different types of health information for browsing and adoption under the different task classification scenarios(ie,correlation assessment).The second part is based on the user's browsing collected by eye tracker.The data determines the health information search context in which the user is located.In addition,whether the user's demographic information and personality,health information literacy,health status and other variables affect the user's information search process is also a factor we consider.This provides a feasible personalized recommendation theory link for search engines and health information websites,that is,by collecting the user's screen eye movement behavior data,predicting the specific task search situation in which they are located,and then personally pushing the situation to be Relevant health type information adopted by most users.This study uses experimental methods,content analysis methods,interview methods,mathematical statistics analysis and other research methods,through the combination of eye movement data obtained in eye tracking experiments and text content viewed by users,summarizing the health information of users in different task situations.The law of assessing the relevance of information during the search;combining the length of each gaze and skimming and the related factors such as demographics and health information literacy in the pre-test questionnaire to determine the health information search context in which the user is located.The results show:Firstly,in different mission scenarios,users tend to judge different types of health information as relevant.Participants in fact-based tasks are trying to figure out the cause of the problem and the potential problem-solving needs,not just browsing the information specified in the task;the exploratory subjects searching for themselves generally believe that based on the patient's health information One category is related to their information needs;the subjects of the exploratory task searched for others are more concerned with all official statements related to a disease or problem and the subjective feelings of the patient;the participants of the empirical task are simultaneously based on The patient's health information and health information based on facts/common senses guide their relevance judgments,judging that the relevant types of information are more diverse,and the proportion of information in several types is more balanced;the type of health information that college students are generally concerned with is Common health issues,hot health issues,and controversial health topics.Secondly,users have different gaze and skimming durations in different mission scenarios.All the browsing durations were the shortest for the fact-finding tasks;the exploratory readings for the searchers who searched for themselves were significantly longer in the reading phase than those who searched for others;the custom tasks and the prolonged glance phase lasted longer.Long-term scanning phase,careful reading duration,long reading period,lengthy reading duration,and short reading period,careful reading duration.Thirdly,according to the different durations of the subjects,combined with gender and health information literacy and other factors,the logistic regression model was used to predict the task search situation of the subjects,the accuracy rate was 85.71%,and the model's 10-fold cross-validation average accuracy rate It is 81.52%.