SBIR Phase II: Big data Analytics Driven Adaptive Learning for STEM Education

这一二期项目将基于自适应学习原理的基于大数据分析的自适应在线学习平台商业化。在线学习系统提供具有实时学习者分析的自适应学习策略。该系统从学习的四个方面整合了学习者分析：多媒体、学习策略、互动性和社交--为科学、技术、工程和数学(STEM)学生提供个性化的学习体验，显著提高学习结果。具有个性化学习策略的自适应学习软件技术平台显示，在线学生参加大专课程的完成率和满意率很高。在此SBIR第二阶段，我们建议开发一个独特的数据驱动的决策支持界面，该界面将为个人和大量学习者提供实时大数据分析。数量、速度、多样性和准确性(大数据的4V)将首先由个别学校的数据收集生成，并有可能聚合来自地区、州甚至国家层面的数据。通过自适应学习平台对学习者轨迹的大数据分析将揭示出可以提高对教育中学习者行为的理解的模式。每种学习策略中与学习者行为相关的自适应学习系统产生的大数据，将为第二阶段移动平台上产品的进一步开发和所提出的方法的有效性提供有价值的见解。学习者分析将为基于统计证据的智能反馈提供基础。所提出的数据驱动的自适应学习决策过程方法是基于对单个学习者的预测变量进行实时互相关统计分析的。提议的方法的实地试验将在参与的高中进行。在高中实地试验期间收集的一组学生的数据将导致发现每种学习策略下的学习者群的学习模式。之前的第一阶段研究已经导致开发了一个分析模型，用于分析全校范围内收集的数据的数量、速度、多样性和准确性。从这些数据中得到的报告被用于决策树的开发和回归分析，以找到可用于学习型企业(学校)改进的相关性(知识发现)。对学生的数据驱动反馈和对教师的群体分析将为提高中小学和大学STEM教育的能力和毕业率提供必要的反馈机制。

This Phase II project will commercialize a big-data analytics-based adaptive online learning platform founded on the principles of adaptive learning. The online learning system provides adaptive learning strategies with real-time learner analytics. The system integrates learner analytics from four dimensional aspects of learning: multi-media, learning strategies, interactivity, and social interaction-- to deliver a personalized learning experience for science, technology, engineering, and math (STEM) students for significant improvement in the learning outcome. The adaptive learning software technology platforms with personalized learning strategies have demonstrated high completion and satisfaction rates for online students taking post-secondary courses. In this SBIR Phase II we propose to develop a unique data driven decision support interface that will result in real-time big data analytics for both individuals and large numbers of learners. The volume, velocity, variety, and veracity (the 4 Vs of big data) will be generated by the collection of data at individual schools first, with the potential to aggregate data from district, state, and even national levels. The big-data analytics of the learner trajectories through the adaptive learning platform will uncover patterns that can improve understanding of learner behavior in education. Big data, generated by the adaptive learning systems related to learner behavior in each learning strategy, will lead to valuable insights on efficacy of the proposed methodology and further development of the product on mobile platforms in the Phase II. The learner analytics will provide the basis for intelligent feedback based on the statistical evidence. The proposed method of data driven decision process for adaptive learning is based on the real- time cross-correlation statistical analysis of the predictor variables for an individual learner. The field trials of the proposed method will be conducted in the participating high schools. Collection of data for a group of students collected during the field trials in high schools will lead to discovery of learning patterns for the clusters of learners in each learning strategy. The previous Phase I research has led to the development of an analytical model for volume, velocity, variety, and veracity of data collected at school-wide level. This reports form this data is used for the development of decision tree and regression analysis to find correlations (knowledge discovery) that can be used for the improvement of learning enterprise (school). The data-driven feedback to students and the group analytics for teachers will provide necessary feedback mechanism for improving competency and graduation rates for STEM education in schools and colleges.