Font Size: a A A

Research And Implementation Of HCI System In Printer Field Based On Natural Language

Posted on:2012-09-14Degree:MasterType:Thesis
Country:ChinaCandidate:P YuFull Text:PDF
GTID:2218330368487753Subject:Electronics and Communications Engineering
Abstract/Summary:PDF Full Text Request
We are now in an information society full of rapid change. The biggest phenomenon is that various kinds of electronic equipment are frequently used in our daily life. However, because of the increasing of quantity and complexity of electronic equipment functions, the operations to those equipments are also become more complicated. Therefore, a friendly and smart interactive way is very important for the user.With the efforts of many researchers, many new interactive ways are constantly appeared, such as touch interaction, voice interaction and so on. However, problems like UI restriction in touch interaction and development bottleneck in voice interaction are still exist to limit their development. For this reason, natural language interaction becomes the hot topic for it's simply operation.A new method is proposed based on the printer field to allow users operating equipment with natural language. The user just input their natural language description of functions, and the system will extract functions from it and execute extracted functions automatically. The designed system can be divided into three parts:knowledge base construction phase, function extraction phase and function execution phase. Ontology is used to construct the knowledge base, and define numerous concepts related to printer filed, which can support the execution of next two phases; A "bridge" way is applied to extract functions contained in users' description, which can avoid errors that most happened in the direct extraction. The improved algorithm based on traditional way play an important role to concept mapping; The ontology is defined with a special structure, mapping scheme generated by it insure the successful execution of functions.A prototype system in printer filed is developed, and more than 1000 concepts are defined in ontology. The test accuracy can reach 89% after testing a large number of natural language inputs, which indicates that the system can realize basic functions of the printer from various manufacturers. The intended target is achieved.
Keywords/Search Tags:Human-Machine Interaction, Functional Information Extraction, Ontology Mapping, Natural Language Processing, Functional Ontology
PDF Full Text Request
Related items