大数据时代网络教育学习成绩预测的研究与实现———以本科公共课程统考英语为例

Research and Implementation of Network Education Results Prediction in Big Data Era: A Case of English Unified Exanination for Online Undergraduates

合适的数据分析技术能使我们借助网络学历教育学生在学习和管理系统中产生的数据和信息， 发现相关规律，进而为网络学历教育教学和管理流程的优化提供有益的决策依据。本文采用数据挖掘中数据分 类C5. 0 决策树方法，通过分析网络学历教育本科学生英语学习及相关信息，实现了对其英语统考成绩的预测。 在分析英语统考前景预测的目标特性后，在SPSS 的Clementine 12. 0 数据挖掘环境中，历经数据提取、数据预处 理、决策树构建和决策树优化等步骤，本研究构建了网络教育本科英语统考成绩的预测模型，并提出了模型实现 方法;同时对模型相关属性的重要性进行了分析，提出了提高网络教育本科学生英语学习水平和统考通过率的 相应策略。

With the development of information society， information storm brought by big data is changing our life， working and thinking style. More and more students are participating in online diploma education. They learn through accessing and using online resources， online homework， interactive discussions and examinations. Such participation has left or generated a giant useful data and information in various types of course management and learning systems. Using appropriate data analysis techniques， these data and information can help us obtain useful knowledge ，find relevant disciplines ， and provide useful basis for decision making， programming all aspects of online learning ，optimizing processes for teaching management， and improving the quality of teaching and changing the design of educational software. All of these are inevitable demands for sustainable development of online education. We analyzed English learning and other relevant information of undergraduate students in online diploma education using data classification technology in data mining， and forecasted the prospects when they took the English unified examination， which was necessary for their graduation. After briefly describing the concepts and relevant theories of data mining， we compared the differences between classification and clustering techniques in data mining and analyzed the characteristics of achieving the goal for forecast. We determined to use C5. 0 decision tree classification in data mining. Using the relevant data of students who took online education in Jiangnan University and had taken the English unified examination as the training data， our research went through four steps ： data retrieving， data preprocessing， decision tree structuring and optimization. Then， in Clementine 12. 0 data mining environment of SPSS， we built a forecasting model of the undergraduate English unified examination. This article also proposes the implementation method of the model， discusses and analyzes the constructed forecasting model and the importance of the related properties ， and proposes appropriate policies about how to improve the level of undergraduate English online education and the throughput rate of the English unified examination.