A DAISY Reader Recommendation System For Visually Impaired People

In 2005,the National Information Standards Organization(NISO)published the DAISY 3.0 standard,which has been endorsed by the American National Standards Institute(ANSI).The DAISY 3.0 standard defines the content and the format of a set of audio books(DTBs).These audio books include XML text,the corresponding audio and other necessary resource files.As opposed to the traditional e-book format,DAISY e-book format which is designed for the visually impaired people and people with printing disabilities has many advantages,such as DAISY allows the visually impaired people to easily browse books.According to the CDPF statistics in 2006,there are about 82.96 million disabled people in China,among which there are 12.33 million visually impaired people,and the number is increasing year by year.Visually impaired people and people with dyslexia do not have the same free reading experience as normal people due to innate or acquired defects.In this context,the China Braille Press hope to develop a DAISY reader for the visually impaired people.Firstly,according to the DAISY 3.0 standard,this paper develops an e-book reader software based on iOS system which can read DAISY 3.0 format.The reader software can navigate by the directory,the sentence and the paragraph,and highlight the corresponding sentence in the audio playback.Secondly,we construct a recommendation system by building a back-end server and realize the online recommendation function of the reader.Initially,we implement the recommendation system using user-based collaborative filtering recommendation algorithm and item-based collaborative filtering recommendation algorithm.Then the paper puts forward the reasonable fusion algorithm based on the mobile end on the analysis of the true data and user feedback obtained by the system and the characteristics of the mobile end environment.Finally,the paper proposes a recommendation algorithm which is applied in mobile application and based on user behavior by considering some characteristics of the visually impaired.It is verified by the experimental results on the real data set that the recommendation algorithm proposed by the paper for the visually impaired population is superior to the traditional recommendation algorithm,which has proved the feasibility of the improved recommender system in enhancing the reading experience of the visually impaired.And its modified version has formally applied in China Braille Press to provide service for the blind.