Utterance Labelling Crowdsourcing Platform Design And Implementation

The task-oriented dialogue system is one of the most popular and useful sub-fields of the intelligent question answering system.The task-oriented dialogue system requires a large amount of data as the basis for training,so it is necessary to process and label these data to be Used in scientific research.However,the amount of data that needs to be annotated is very large.If only scientific researchers are required to annotate the data,the workload is very large.In a team,it is impossible for all staff members to have the same skills in all professional skills.Especially in many rural areas,it is not easy for us to find staff with professional skills.In recent years,with the further development of the Internet,almost everyone browses various website information almost every day,because everyone has different interests and hobbies,there must be corresponding differences in the areas they are good at.we can mark and comment on the massive data on the Internet,and work in the form of crowdsourcing.Then it can make it easier for researchers to obtain valuable information and it will be more convenient for researchers to conduct corpus analysis and carry out machine learning research.This project was developed based on this demand and provided a platform for users to browse,label and comment on information.At the same time,when users mark and comment information,data mining and machine learning algorithms are also used to recommend different data for users according to their different problems.This system is implemented using a 3-tier client server approach,with a backend database which stores the rules generated from the f-p growth algorithm and also knowledge base from the workers,datasets and slot actions extracted from the attributes of the datasets.The steps that would take to develop the software include gathering of information,building the database and use cases,building the application using PHP,CSS,HTML,JavaScript,MySQL and java.This project was ran on a Freemarker Apache Server.In other to ensure the implementation of our system that has a user-friendly screen,we made use of the latest tools and technologies that is invoke in today's software world.At present,the system has been tested,all the test results meet the requirements,and it is ready for use.The end product delivered will be a used for companies in Nigeria.