| With the development of Embedded technology and the interactive demands of people,there emerged a large number of input methods which based on embedded. These inputmethods provide convenience for the majority of users to interact with embedded systems.But, most of the embedded Chinese input methods are the technology and patents offoreign enterprises. In addition, due to the constraints of the hardware and softwareconditions embedded input methods have poor performances on recognition rate, responsetime and easy-to-use. However, the development of cloud computing takes a fatal chance tochange this situation. Free installation, mass thesaurus and high recognition rate of thecloud input method made a further research of input method. This thesis discusses how touse the cloud computing and embedded technology to design input method and to form anembedded input method architecture in cloud computing environment. So that, it couldbetter meet the information exchange demands of users in daily life.This thesis mainly researched the design of software and hardware of embeddedterminal device and the construction of input method architecture. Firstly, we introducedthe development process of input method, and analyzed domestic and foreign popularembedded input method from keyboard input, handwriting input and voice input fields, andextracted the network input method which suitable for embedded cloud computing platform.Secondly, we introduced the design of the embedded terminal which include hardware andsoftware. The hardware part include embedded hardware terminal framework design andthe selection of embedded development board, CPU and LCD touch screen. The softwarepart includes the selection designation of embedded system and graphical user interface.Thirdly, we divided the input method system into three subsystems which include keyboardinput subsystem, handwriting input subsystem, voice input subsystem and design themrespectively. Finally we proposed an optimization method of cloud input method whichbased on mining and multi-attribute matching of user interest vocabulary. We introducedefficient parallel frequent item-sets data mining method of massive text databases to findthe interests vocabulary of users. Then use the multi-attribute bilateral matching methodwhich considering attributes weights to calculate the overall matching degree. At last, wemake the optional matching result as the candidate words which fulfill the personalitydemands and recommend the candidate words to users. All of this enables users to easilyand efficiently input, so as to meet the special information demand of users. |
