【作者】 王哲；

【导师】 奚建清；

【作者基本信息】 华南理工大学 ， 软件工程， 2024， 硕士

【摘要】 计算机行业的蓬勃发展吸引了众多非科班出身的人员涌入计算机行列,这些非科班出身的普通用户虽然思维灵活,不受传统学科的束缚,能够提出创新性的想法,但是由于他们对计算机知识的学习层次不一,有的完全是零基础,有的有一定的基础,还有的已经具备一定的专业知识,往往需要花费大量的时间和精力系统地学习计算机基础知识和技能。为了让普通用户快速上手体验开发,笔者实验室研发了一款面向中文的业务命令语言——帮语。但是目前帮语缺乏对物流场景的支持,不能很好地满足物流领域专业人员的需求,不能有效地帮助他们完成物流管理中的各种任务。此外,帮语目前只能手动地对句型库中已有的句型进行填充才能完成语句的编写,无法让用户自行编写语句,在使用方面缺乏一定的灵活性。为了解决上述问题,本文开展了以下研究工作:(1)针对目前帮语缺乏对物流场景支持的问题,本人通过文献阅读、Chat GPT对话交流及对实际物流场景的观察和调研等方式对物流行业进行深入分析,描述了物流场景的常见业务,使帮语能够为物流领域专业人员提供更加专业和实用的帮助,根据帮语固有的句型功能类型及语义域类型提出了文本分类和序列标注的融合模型,来确定业务语句所属的句型类别以及业务语句含有的语义信息。(2)为了提高帮语使用方面的灵活性和用户体验,设计业务语句向帮语语句转换的语句转换架构完成语句的单句转换。首先,基于设计的融合模型获取句型功能类别和语义信息,然后根据句型功能类型和语义信息与句型库中句型进行匹配,根据匹配结果设计不同的解决方案以得到目标句型,最后将语义值填充到目标句型中;基于语句转换架构完成业务描述到分析剧本的转换,使普通用户可以自行编写语句,不再拘泥于既定的句型,提高帮语使用的灵活性。(3)利用设计好的语句转换架构对物流场景的业务描述进行转换,并对转换形成后的分析剧本进行模拟,从转换的准确率和执行正确率两方面验证了语句转换架构的合理性和有效性,也证明了帮语能够满足物流场景的需要。更多还原

【Abstract】 The booming development of the computer industry has attracted many people from noncomputer backgrounds to join the computer industry.Although these non-computer backgrounds ordinary users have flexible thinking and are not bound by traditional disciplines,they can propose innovative ideas.However,due to their varying levels of learning computer knowledge,some have no foundation,some have a certain foundation,and some already have a certain level of professional knowledge,they often need to spend a lot of time and energy systematically learning computer basic knowledge and skills.In order to enable ordinary users to quickly get started with development,my laboratory has developed a Chinese oriented business command language band-language.However,currently band-language lacks support for logistics scenarios,which cannot meet the needs of logistics professionals and effectively assist them in completing various tasks in logistics management.In addition,currently,bandlanguage can only complete sentence writing by filling in existing sentence patterns in the sentence pattern library,and users cannot write sentences themselves,which lacks a certain degree of flexibility in use.To address the aforementioned issues,this article conducted the following research work:Firstly,in response to the current lack of support for logistics scenarios in band-language,I conducted in-depth analysis of the logistics industry through literature review,Chat GPT dialogue and communication,as well as observation and research on actual logistics scenarios.I described common businesses in logistics scenarios,enabling band-language to provide more professional and practical assistance to professionals in the logistics field.Based on the inherent sentence function types and semantic domain types of band-language,I proposed a fusion model of text classification and sequence annotation to determine the sentence type and semantic information contained in business statements.Secondly,in order to improve the flexibility and user experience of using band-language,design a statement conversion architecture for converting business statements into bandlanguage to complete the single sentence conversion of statements.Initially,based on the designed fusion model,the functional categories and semantic information of sentence patterns are obtained.Then,according to the functional categories and semantic information of sentence patterns,they are matched with sentence patterns in the sentence pattern library.Based on the matching results,different solutions are designed to obtain the target sentence pattern;Eventually,fill the semantic values into the target sentence pattern.Based on the statement conversion architecture,complete the conversion from business description to analysis draft,allowing ordinary users to write their own statements without being bound by established sentence structures,improving the flexibility of using band-language.Finally,using the designed statement transformation architecture to transform the business description of logistics scenarios,and simulating the analysis draft formed after the transformation,the rationality and effectiveness of the statement transformation architecture were verified from the perspectives of conversion accuracy and execution accuracy,and it was also proved that the band-language can meet the needs of logistics scenarios.更多还原