Task Assignment And Quality Control Of Crowdsourcing-Based Web Accessibility Evaluation System

The Internet era affacts people's lives,has become an integral part of our daily life.However,there are still a considerable number of special groups in today's society,such as the visually handicapped and the hearing impaired.They can not access online information like normal people and enjoy the convenience brought by the Internet.The rapid development of the Internet increased the gap between the disabled and the normal people.In order to make people with disabilities more convenient to visit the website,the site design should meet the web accessibility standards.To assess the accessibility of the site,it is not enough to simply use the automated detection tools because many of the accessibility checkpoints require manual participation.Crowdsourcing-Based Web Accessibility Evaluation System is used to assist users with accessibility detection.The system uses the wisdom of users in the Internet,so that more people can participate in the detection of web accessibility.However,the volunteers' professional background,ability and so on are quite different.Simple random task assignment can't use the characteristics of volunteers and assign them appropriate tasks.Therefore,it is necessary to find a new method that use the characteristics of volunteers and tasks.In addition,the results of a volunteer often do not have the reliability,therefore,a task usually requires multiple volunteers to complete together.How to control the quality of task results is a problem to be considered.In this paper,we propose a task assignment method based on genetic algorithm and confidence,which assigns the detection tasks according to the user's characteristics.We extract the user characteristics of correct rate,error rate and abandonment rate of different detection rules and use genetic algorithm to calculate the weight of each feature in the task assignment.At last we use the confidence to measure the differenct between the new task assignment and random task assignment method.In the aspect of task quality control,this paper mainly lists the task quality control methods used in Crowdsourcing-Based Web Accessibility Evaluation System,such as eliminating malicious detection users.we propose a method of task quality control based on weighted trustworthiness for the final decision method of crowdsourcing tasks.This method combines the reliability of the volunteer and the confidence of the task results to calculate the value of the credibility.After getting the value we should find a method to minimize the error in the result so that we can control the quality of the crowdsourcing tasks.Finally,this paper extracts the data of Crowdsourcing-Based Web Accessibility Evaluation System and validates the method proposed in this paper.It is proved that the new task assignment method and the quality control method are better than the original methods.