A Study On Related Problems Of Chinese Input Method

In recent years, with the rapid development of the computer science technology and Internet, the Chinese input is becoming more and more significant in people’s daily life. Because of the speciality of the Chinese coding, we cannot directly input Chinese into the Computers, we have to only depend on the Chinese input method’s coding scheme. Now some of the most popular Chinese input methods include Chinese Pinyin Input Method, five stroke input method for Chinese and so on. Compared with other Input Methods, Pinyin Input Method has some advantages such as simple, easily learning etc, so currently Pinyin input method has become one of the most common used input method for the Chinese input. However, Pinyin input method itself has many disadvantages, for example, it is complex for the users to select the candidates words; Once the user wrongly typed, it would be very troublesome for them to modify towards the correct answer; Besides, the whole sentence input method also can not provide the users a satisfiable performance. Recently, with the quick development of the technology of mobile platform, more and more people are beginning concentrated on the topic that how to transplant the input method of the PC platform to that of the mobile platform and how to optimize the input method of the mobile platform, as we all know, the size of the devices of the mobile platform such as the mobile phone is limited, thus declining the efficiency of the people’s input. Meanwhile, the accuracy of current whole sentence input is still low, which is far to reach the requirement that input the whole sentence fluently.Based on these situations, in this thesis, we mainly propose a new intelligent Chi-nese spelling checking algorithm, and how to make use of the statistical machine trans-lation framework to improve the accuracy of the Chinese Pinyin-to-Character(PTC) conversion problem, especially for the whole sentence’s PTC conversion. Finally, we try to present a new Chinese Input Method’s UI design technique and optimize the can- didate letter predication problem based on this UI, thus effectively reducing the number of the users’key stroke, improving the accuracy of the users’input and saving the time.The innovating points of this thesis lie in that:by introducing the spell checking technology to automatically check the Chinese text, we can remind users potential spelling errors, thus saving the user’s time;by introducing the framework of statistical machine translation to optimize the Input Method technique, we improve the decoding accuracy of the PTC problem, especially the whole sentence’s accuracy;aiming at the mobile platform, we make a special UI design and algorithms to predicate the possible candidate letters, we fully incorporate the characteristics of the mobile platform into the Input Method and improve the speed and accuracy of users’input.