Enhancing Teaching and Learning through Educational Data Min(4)

to timed progress and self-report of derived from past activities of all students, such as “like” engagement, CyGaMEs captures indicators, assessment results, and correlations between behaviors the player uses during play.

Reese et al. (in press) showed that this student activity and assessment results. The student chooses behavior data exposed a prototypical a resource to work with, and his or her interactions with it are “moment of learning” that was confirmed used to continuously update the system’s model of how by the timed progress report. Research

using the flow data to determine how user much he or she knows about population genetics. After the experience interacts with learning is student has worked with the resource, the dashboard shows ongoing. updated ratings for each population genetics learning

resource; these ratings indicate how much of the unit content

the student has not yet mastered is covered by each resource. At any time, the student may

choose to take an online practice assessment for the population genetics unit. Student responses to this assessment give the system—and the student—an even better idea of what he or she has already mastered, how helpful different resources have been in achieving that mastery, and what still needs to be addressed. The teacher and the institution have access to the online learning data,

another example of data use for “sensing” student learning and engagement is described in the

sidebar on the moment of learning and illustrates how using detailed behavior data can pinpoint cognitive events.

The increased ability to use data in these ways is due in part to developments in several fields of computer science and statistics. To support the understanding of what kinds of analyses are possible, the next section defines educational data mining, learning analytics, and visual data analytics, and describes the techniques they use to answer questions relevant to teaching and learning.

美国教育部教育数据挖掘概述文档

Data Mining and Analytics: The Research Base

Using data for making decisions is not new; companies use complex computations on customer data for

business intelligence or analytics. Business intelligence techniques can discern historical patterns and trends from data and can create models that predict future trends and patterns. Analytics, broadly defined, comprises applied techniques from computer science, mathematics, and statistics for extracting usable information from very large datasets.

An early example of using data to explore online behavior is

Web analytics using tools that log and report Web page

visits, countries or domains where the visit was from, and the

links that were clicked through. Web analytics are still used

to understand and improve how people use the Web, but

companies now have developed more sophisticated

techniques to track more complex user interactions with their

websites. Examples of such tracking include changes in

history for predicting likely Web pages of interest, and

changes in game players’ habits over time. Across the Web,

social actions, such as bookmarking to social sites, posting to

Twitter or blogs, and commenting on stories can be tracked

and analyzed. Unstructured Data and Machine Learning Data are often put into a structured format, as in a relational database. Structured data are easy for computers to manipulate. In contrast, unstructured data have a semantic structure that is difficult to discern computationally (as in text or image analysis) without human aid. As a simple example, an email message has some structured parts—To, From, and Date Sent— and some unstructured parts—the Subject and the Body. Machine learning approaches to data mining deal with unstructured data, finding patterns and regularities in the

data or extracting semantically meaningful

information.

Analyzing these new logged events requires new techniques

to work with unstructured text and image data, data from

multiple sources, and vast amounts of data (“big data”). Big data does not have a fixed size; any number assigned to define it would change as computing technology advances to handle more data. So “big data” is defined relative to current or “typical” capabilities. For example, Manyika et al. (2011) defines big data as “Datasets whose size is beyond the ability of typical database software tools to capture, store, manage, and analyze.” Big data captured from users’ online

behaviors enables algorithms to infer the users’ knowledge, intentions, and interests and to create models for predicting future behavior and interest.