随着人工智能时代的到来，编程教学在中小学中愈发受到重视，程序设计的教学内容已经逐步进入中小学教学与考试体系中。由于编程辅导系统能够弥补课时少和师资短缺等不足，让学习者在编程过程中及时获得有效的指导和反馈，编程辅导系统成为了程序设计教学在中小学落地的重要教学工具。然而，学习者不当利用系统辅助功能，出现了滥用或回避使用系统帮助的现象，使得编程辅导系统在真实教学应用中收效甚微。

为解决这一问题，本研究设计开发了具有求助行为即时调节的编程辅导系统MixToyIT。围绕MixToyIT与学习者求助行为分析，本研究开展了三项研究内容。第一，本研究设计开发了具有求助行为即时调节功能的编程辅导系统。本研究基于自我调节学习理论和智能辅导内循环过程，总结了求助行为即时调节的理论模型，用以指导系统中的求助行为的即时调节。基于该模型，本研究对MixToy图形化编程软件添加了用户管理、教学管理、编程辅导和求助调节等模块，开发了在线的编程辅导系统MixToyIT。系统实现了对学习者不良求助进行即时反馈的功能，用以改善学习者在编程辅导系统中的编程学习过程。并且本研究面向学生和教师开展问卷调查，以考察用户对该系统的技术可接受度。第二，为了验证求助行为即时调节功能的应用效果，本研究设计了准实验研究探究来求助行为即时调节对学习成绩、学习动机和使用评价上的影响。其中实验组学生有求助行为即时调节而对照组无求助行为即时调节。研究结果表明实验组的成绩和系统使用评价显著高于对照组，并且实验组中不同性别、编程经验、前测水平的学生在成绩上没有显著差异，即求助行为即时调节功能具有普适性。第三，本研究分析和挖掘了MixToyIT中学习者的求助过程。本研究基于日志数据提取行为特征，并对行为特征进行差异性分析，结果表明在求助频次和持续编程上实验组与对照组有显著差异，并且发现这两个行为特征和成绩呈显著正相关关系。此外，本研究利用离散马尔可夫模型和Kmeans聚类挖掘得到积极求助策略、依赖性策略、尝试求助策略、直接完成策略、独立解决策略和应付性策略六类求助策略，并发现求助行为即时调节对依赖性策略和应付性策略两个不良的求助策略有积极影响。

总之，本研究基于求助调节理论模型，设计开发了具有求助行为即时调节功能的编程辅导系统MixToyIT，并开展了系列研究，得到的研究结果是：MixToyIT具有良好的用户体验和普适性，其中的求助行为即时调节功能对学习者成绩、持续编程和求助行为有良好的促进作用。这一结果扩展了编程辅导系统设计开发方向，对探索求助行为即时调节如何影响学习者求助过程和编程表现具有重要意义。

With the development of artificial intelligence technology, programming teaching has received more and more attention in primary and secondary schools, and computer programming has gradually entered the curriculum of primary and secondary schools. Due to a fact that Intelligent Programming Tutors can compensate for the shortage of class hours and teaching staff, allowing learners to receive timely and effective guidance and feedback during the programming process, Intelligent Programming Tutors have become an important teaching tool for programming teaching in primary and secondary schools. However, learners misuse system assistance functions, such as abusing or avoiding the help of Intelligent Programming Tutors. As a result, Intelligent Programming Tutors have little effect in real teaching applications.

To solve this problem, an Intelligent Programming Tutors with help-seeking behavior regulation, MixToy Intelligent Tutors (MixToyIT) was designed and developed. Focusing on MixToyIT and learners’ help behaviors analysis, three research contents were carried out in this study. First, this study designed and developed an Intelligent Programming Tutor with help seeking behavior regulation. Based on the theory of self-regulated learning and the internal circulation process of intelligent tutoring, a theoretical model of help-seeking behavior regulation was summarized to guide the regulation of help-seeking behavior in the system. Based on this model, this study added modules such as user management, teaching management, programming guidance, and help seeking behavior adjustment to the MixToy block-based programming software, and developed an online Intelligent Programming Tutor. The system has implemented the function of providing feedback on learners' help seeking behavior, in order to improve learners’performance in Intelligent Programming Tutor. And this study conducted a questionnaire survey for students and teachers to examine the user's technical acceptance of MixToyIt. Second, in order to verify effect of help-seeking regulation in MixToyIT, this study designed a quasi-experimental study to explore the impact of help-seeking regulation on learners' performance, user's technical acceptance and motivation. Among them, the students in the experimental group had help-seeking regulation, while the control group did not have help-seeking regulation. The results showed that the experimental group performed significantly higher than the control group. And there was no significant difference in the performance of students of different genders, programming experience, and pre-test level, which means the help-seeking regulation function was universal. The third part of this study analyzes and mines the learner behavior process. In this study, the behavioral characteristics were extracted based on log data. Differences of behavioral characteristics were analyzed, and the results showed that there were significant differences between the experimental group and the control group in terms of help frequency and continuous programming. Moreover, there was a significant positive correlation between these two behavioral traits and achievements. In addition, Discrete Markov Model and Kmeans cluster were used to mine six types of help-seeking strategies including positive help strategy, dependency strategy, attempt to help strategy, direct completion strategy, independent solution strategy and coping strategy. Morover it is found that help-seeking regulation has a positive effect on the two bad strategies of dependency strategy and coping strategy.

In summary, this study designed and developed an Intelligent Programming Tutor named MixToyIT which has a function of help seeking behavior regulation, and this study conducted a quasi-experimental study to verify it. The results of this study showed that MixToyIT has a good user acceptance and universality, and immediate regulation of help seeking behavior has a good promoting effect on learner performance, continuous programming, and help seeking behavior. This result expands the direction of Intelligent Programming Tutors design and development, and is of great significance in exploring how regulation of help-seeking behavior affects learners' help-seeking process and programming performance.