Conference PaperPDF Available

Recommendation of Personalized Learning Materials based on Learning History and Campus Life Sensing

Authors:
  • Kyushu University(home institution) and University of Michigan(Visiting)

Abstract and Figures

Thanks to the widespread of ICT environments not only in social life but also in the educational field, personalized learning has become a real possibility in recent years; it is expected to provide adaptive support to learners based on their situations. A typical approach is to recommend personalized learning materials based on the learning progress or understanding level of learners. Effective recommendations will maintain learners' motivation and help them overcome weaknesses. However, too many and/or too frequent recommendations sometimes interfere with learners' interests. Therefore, it is important to consider the timing of when recommendation information should be sent to learners. In other words, we assume that timely recommendations will encourage learners' motivation more than greedy strategies. In our study, we proposed a new strategy to support personalized learning. Our approach collaborates with activity sensing during campus life and automatically detects the timing of recommendations. Moreover, our recommendations provide a short summary of learning materials, which enhances learners' previews compared with the original materials. In this paper, we introduce the configuration of the proposed system, followed by a report of preliminary experimental results and a mention of future works.
Content may be subject to copyright.
Companion Proceedings 10th International Conference on Learning Analytics & Knowledge (LAK20)
Creative Commons License, Attribution - NonCommercial-NoDerivs 3.0 Unported (CC BY-NC-ND 3.0)
Recommendation of Personalized Learning Materials based on
Learning History and Campus Life Sensing
Keita Nakayama1*, Atsushi Shimada2*, Tsubasa Minematsu3*,
Masanori Yamada4*, Rin-ichiro Taniguchi5*
*Kyushu University, Japan
1nakayama@limu.ait.kyushu-u.ac.jp, 2atsushi@ait.kyushu-u.ac.jp,
3minematsu@limu.ait.kyushu-u.ac.jp, 4mark@mark-lab.net, 5rin@kyudai.jp
ABSTRACT: Thanks to the widespread of ICT environments not only in social life but also in the
educational field, personalized learning has become a real possibility in recent years; it is
expected to provide adaptive support to learners based on their situations. A typical approach
is to recommend personalized learning materials based on the learning progress or
understanding level of learners. Effective recommendations will maintain learners' motivation
and help them overcome weaknesses. However, too many and/or too frequent
recommendations sometimes interfere with learners' interests. Therefore, it is important to
consider the timing of when recommendation information should be sent to learners. In other
words, we assume that timely recommendations will encourage learners' motivation more
than greedy strategies. In our study, we proposed a new strategy to support personalized
learning. Our approach collaborates with activity sensing during campus life and automatically
detects the timing of recommendations. Moreover, our recommendations provide a short
summary of learning materials, which enhances learners' previews compared with the original
materials. In this paper, we introduce the configuration of the proposed system, followed by
a report of preliminary experimental results and a mention of future works.
Keywords: recommendation, adaptive learning, learning history, activity sensing
1 INTRODUCTION
With the development of information technology, various ICTs have been introduced into the
educational and learning environment, along with increasing expectations for the realization of
personalized learning environments that can recommend appropriate learning materials (LMs) based
on the learner's history of learning and provide support to enhance the learning effect (Hwang et al,
2017; Yin & Hwang, 2018). On the other hand, the teaching style is still the traditional face-to-face and
multi-person simultaneous lecture style. For this reason, it is difficult to conduct lectures based on the
learning progress and understanding level of individual learners. Furthermore, with the diversification
of learning methods and lifestyles, providing the same LMs to all learners is not appropriate. Therefore,
it is necessary to realize adaptive learning support that matches individual situations (Truong, 2016).
One approach to learning support based on the individual learning situation is the personalized
recommendation of digital LM based on the learning situation. Many traditional methods recommend
the use of LMs in a lecture or self-learning (e.g., Lan & Baraniuk, 2016; Wan & Niu, 2018). Thus, it is
assumed that the learner is learning at the time of recommendation. To provide further learning
support, it is important to use not only conventional methods but also to promote learning in daily life
even outside of learning time. To encourage learning outside of learning time, it is important for the
649
Companion Proceedings 10th International Conference on Learning Analytics & Knowledge (LAK20)
Creative Commons License, Attribution - NonCommercial-NoDerivs 3.0 Unported (CC BY-NC-ND 3.0)
system to consider the timing to recommend (Pielot et al, 2015). With the spread of smartphones, it
has become possible to make use of data measuring daily life and these data has become using in
recommendation method (Okoshi et al., 2019). Using these technologies, it will be possible to
promote learning appropriately in daily life outside of study hours.
This study aims to provide learning support that effectively occupies learners' free time by individually
recommending LMs based on learning history and campus life sensing, using university education as
a test environment. In this study, we propose a learning support system that recommending lecture
slides as digital LMs composed of suitable contents and quantities for each learner at appropriate
times by integrating the three methods of detection by a smartphone of free time when the learner
may study, recommendation based on learning history and learner's activity and automatic
summarization of digital LMs. In the following section of this paper, we introduce the configuration of
the proposed system and each method followed by a report of preliminary experimental results.
2 PERSONALIZED RECOMMENDATION SYSTEM BASED ON LEARNING
HISTORY AND CAMPUS LIFE SENSING
2.1 System overview
In this study, we propose a learning support system that recommends lecture slides as digital LMs
composed of suitable content and amount for each learner at an appropriate time based on the
learner's current activity and learning history. The proposed system takes three steps. Step 1 is that
learning time is detected by a sensor in a learner's smartphone. Step 2 is that LM is determined based
on the learner's learning history. Step 3 is that the LM determined in the step 2 is recommended to
the learner via email at the time detected by the step 1.
Figure 1 shows the configuration of the proposed system. Moodle (Flanagan & Ogata, 2018) is a
learning management system. BookRoll (Flanagan & Ogata, 2018) is an e-Book system which has
various data such as LMs, learning logs, etc. Teachers and students access BookRoll via Moodle and
browse LMs. This system detects the learner's free time to study and recommends LMs based on the
learner's learning history. The systems for realizing this function are an application measuring the
learner's activity and a campus activity server (CAS). The application is installed on the learner's
smartphone and measures the learner's activity using the sensor installed in the device. Based on the
measured data, the application detects whether the learner is ready to learn and notifies the CAS. The
CAS manages recommendations to learners. The CAS determines the recommended LM and timing
from the learning logs recorded in BookRoll and the information notified by the application and then
sends an email. This is to encourage learning at times when the learner does not intend to study,
which can be a significant burden on the learner. It is therefore desirable that the recommended LMs
enable the learning of important contents in a short period of time. For this reason, this study
recommends LMs with summarized contents. Shimada et al (2016) proposed automatic
summarization method based on the advance organizer theory (Ausubel, 1960), and suggested that
short summaries may have enhanced students’ motivation to preview the material. Therefore, it is
considered that the summarized version of LM has less burden on the learner. Our summarization
system is inspired by (Shimada et al., 2016) and can summarize LMs registered in BookRoll. Details of
the application measuring learner's activity, CAS, and summarization system are described in the
following sections.
650
Companion Proceedings 10th International Conference on Learning Analytics & Knowledge (LAK20)
Creative Commons License, Attribution - NonCommercial-NoDerivs 3.0 Unported (CC BY-NC-ND 3.0)
2.2 Learning time detection
In this study, we recommend LMs via email and encourage learning during the learner's free time. The
reason for measuring learner's activity is that the effect of recommendation timing on the results has
been clarified in research such as ubiquitous computing (Pielot et al, 2015), and the recommendation
effect is higher when the user's situation is taken into account (Okoshi et al., 2019). We have
developed an application to measure learner activity and detect the free time when the learner may
study. Most learners usually always carry smartphones, and the movement of smartphones is thought
to be influenced by the learner's activity. Therefore, the activity of the learner is measured using the
sensor installed in the smartphone. In this study, we use an Android phone and measure the learner's
activity using the acceleration sensor. The Android phone can measure acceleration along three axes.
In this study, it is assumed that the learner may study when the terminal is not moving so hard. This
is because, it is considered that the timing when the learner may study (such as not moving, on the
bus, etc.) is when the terminal is in a stable state (Li et al, 2013). When the measured three-axis
acceleration does not exceed a certain threshold value for a certain period of time, the developed
application detects that the learner is ready to learn and notifies the CAS.
2.3 Recommendation based on learning history and learner activity
The campus activity server (CAS) recommends LMs to each learner based on the learning history
collected from the BookRoll learning logs and activity status obtained from the application. The
recommendation method takes four steps. First, the CAS receives notification from the application
that the learner may study. Second, the CAS checks the recommendation history logs and decides
whether to recommend. Third, the CAS decides which LMs to recommend based on learning history
logs stored in BookRoll. Finally, the CAS sends an email recommending LMs. This server manages
recommendations based on two types of logs: One is a history of recommendation to learners, while
the other is the learner's browsing history for recommended LMs. First, each log is explained, and then
the recommendation method is explained.
The record of a recommendation to a learner is recorded on the server when an email is sent. The
recorded information includes the ID of the recommended user, the ID of the recommended LM, the
recommended time, and the number of pages of recommended LM. The logs of a learner's browsing
Figure 1: System configuration
Summarization System (SS)
Summarize learning materials in
BookRoll and output summary
materials
App in Mobile Device
Measure learner’s activity
Detect time when learning is possible
Receive email from CAS
Teach er Operate SS via Moodle
Download summary materials from
the SS and register with BookRoll
Learning Management
System (LMS)
Moodle
Learner
Learning materials recommended
via email from C AS
Learning b y using BookRoll
Campus Activity Server (CAS)
Analyze learning activity log
Recommend learning materials based
on learning situation and learner’s
activity by email
e-Book system
Having several data in databases
e-book of learning materials
User’s learning activity logs
BookRoll
651
Companion Proceedings 10th International Conference on Learning Analytics & Knowledge (LAK20)
Creative Commons License, Attribution - NonCommercial-NoDerivs 3.0 Unported (CC BY-NC-ND 3.0)
history are a collection of learning logs for the LMs stored in BookRoll. Figure 2 shows samples of
learning logs in BookRoll. The learning logs record operations such as page transitions for digital LMs.
The CAS counts the browsing time of each page of the recommended LM when deciding which LM to
recommend. The log of a learner's browsing history is shown in Figure 3. In Figure 3, "Browsing time"
indicates the number of seconds
T
of accumulated browsing time after recommendation, while "Read
flag" indicates whether learning has been completed. When the contents of page
i
of LM has been
learned, the value of "Read flag"
𝑓
$
is
1
; otherwise, it is
0
.
Next, the recommendation method is described. First, the server is notified by the application. When
the server receives the notification, it checks the recommendation history logs and decides whether
to recommend. In this study, the current time is compared with the last recommended time, and the
LM is recommended to the learner if it has not been recommended for a certain period of time. When
recommending to the learner, the LM to be recommended is determined from the logs of the learner's
browsing history. In this paper, as an initial stage of study, the server recommends LMs in a
predetermined order. When learning with the LM is completed after the recommendation, the system
recommends the next item in the order of LM. In this paper, when each page of the LM is viewed (
𝑇$>
0
), it is assumed that the content of the page is learned (
𝑓
$= 1
). If all the pages of the recommended
LM are not viewed, the LM recommended the last time is recommended again. Finally, the LM
determined by the above method is recommended by email.
2.4 Summarization of learning materials
In this section, we provide an overview of the automatic summarization system (Shimada et al., 2016).
This system was designed to produce a set of lecture slides. The purpose of slide summarization is to
select a subset of pages that maximizes the importance of content under a given condition (in this
case, browsing time). To achieve this, the system selects the important pages without losing the
overall narrative of the lecture. In this section, we will give an overview of the summarization method.
First, lecture material is analyzed to extract important visual and textual features from each page. In
terms of visual importance, the number of objects such as text, figures, formulas, etc. in each material
is estimated, using a background subtraction technique and an inter-frame difference technique.
Word importance is estimated using the TF-IDF method. Furthermore, a teacher specifies the desired
browsing time for students to study each page. The visual, textual, and temporal features determined
by these methods are then combined to generate an importance score. Finally, an optimal subset of
pages is selected, which maximizes the importance score for learning in a short time.
3 EXPERIMENT
We conducted preliminary experiments to confirm the performance of the proposed system and
students' reactions to recommendations. The preliminary experiment was conducted beforehand in
User ID
e-book ID
book title
Page
Operation
Last_read time
xxxxxxxx
oooooooo
5
NEXT
2019
-09-01 11:26:15
xxxxxxxx
oooooooo
6
NEXT
2019
-09-01 11:26:50
xxxxxxxx
oooooooo
7
PREV
2019
-09-01 11:27:45
xxxxxxxx
oooooooo
6
NEXT
2019
-09-01 11:28:35
xxxxxxxx
oooooooo
7
CLOSE
2019
-09-01 11:30:15
Figure 2: Learning logs stored in BookRoll
User ID
e
-book ID
Page
Browsing_time
Read_flg
Last_read_time
xxxxxxxx
ooooooooo
1
100
1
2019
-09-01 **:**:**
xxxxxxxx
ooooooooo
2
120
1
2019
-09-01 **:**:**
xxxxxxxx
ooooooooo
3
60
1
2019
-09-01 **:**:**
xxxxxxxx
ooooooooo
4
0
0
xxxxxxxx
ooooooooo
5
0
0
Figure 3: learner's browsing history logs
652
Companion Proceedings 10th International Conference on Learning Analytics & Knowledge (LAK20)
Creative Commons License, Attribution - NonCommercial-NoDerivs 3.0 Unported (CC BY-NC-ND 3.0)
about 10 hours for 4 university students on campus. In this experiment, if the acceleration of the three
axes did not change more than
5 𝑚 𝑠-
for about 10 minutes, the CAS was notified that it was free
time for learning with the application. Based on the notification, an email was sent recommending
LMs with an interval of at least one hour. The email contains the title of the recommended LM and a
link to Moodle. The recommended LMs are summaries of the materials that will be used in the data
science course in the second semester of 2019/2020 at our university. The recommended LMs
consisted of 8-16 pages, 20% of the original LMs. In this experiment, the response to the
recommended email was arbitrary considering the actual usage situation.
Figure 4 shows a sample of acceleration change of a subject. The change in acceleration in the Figure
4 shows that part "A" is running, part "B" is on the bus, and part "C" is walking. From this result, it was
confirmed that various human activities can be observed by the application. From this measurement
result, the application notified the server that the learner's free time was between 8 and 9 a.m. Note
that with the proposed method, LMs may be recommended if the status of Part B persists for a certain
period of time. In this way, it was confirmed that the developed application could recommend LMs at
a time that seemed to be the learner's free time.
As a result of the experiment, the LMs were recommended 35 times for the learners, and the
recommended LMs were accessed 23 times. The average time from receiving the email to accessing
the recommended digital LMs was 12 minutes 4 seconds. From this result, it can be seen that the LMs
were accessed in a relatively short time after receiving the email, and the recommendation was
conducted when the learner was able to study. The average browsing time for each recommended
LM was 3 minutes 41 seconds. From this result, we can see that each LM can be browsed in a short
time as we intended. Figure 5 visualizes the scatter plots of the access timing vs. the browsing time.
Note that the same LM was recommended multiple times, so that the number of recommendation
was higher than the number of LMs prepared in advance. The x-axis represents the time from when
the email was sent to the learner until the digital LM was accessed, and the y-axis is browsing time
per access to the LM. From Figure 5, we can see there is no correlation between the access timing and
browsing time. This is an expected result because the proposed method sent the recommendation
information when the subject was estimated to have time, and whether he/she accessed to the LM is
completely up to the subject. From the results that the subjects accessed the materials less than 600
seconds in most cases, the timing of recommendation was almost suitable to the learners, and the
learner studied using the free time. In the future, we plan to experiment with more learners and verify
the educational usefulness of the proposed system.
Figure 5: The distribution of the access timing vs. the
browsing period
Figure 4: Sample of sensing data
A B C Notification timeNotification time
653
Companion Proceedings 10th International Conference on Learning Analytics & Knowledge (LAK20)
Creative Commons License, Attribution - NonCommercial-NoDerivs 3.0 Unported (CC BY-NC-ND 3.0)
4 CONCLUSION
In this paper, we proposed a learning support system by recommending learning materials based on
learning history and campus life sensing. The proposed system recommends learning materials
suitable for the learner based on the learning history during the learner's free time detected by the
smartphone. In the previous experiment, a small number of subjects used the system, and we
confirmed the behavior of the system and the response to recommendations from learners.
As future work, we plan to conduct long-term experiments on more subjects using the proposed
system, based on the results of this preliminary experiment. In this paper, as an initial stage of study,
learning materials were recommended in a predetermined order based on learning history. We will
also consider how to recommend learning materials using other types of data such as student
knowledge information, quiz results recorded in Moodle, and so on.
ACKNOWLEDGEMENTS
This work was supported by JST AIP Grant Number JPMJCR19U1, iLDi Grand Challenge 2018-#1, and
JSPS KAKENHI Grand Number JP18H04125, Japan.
REFERENCES
Ausubel, D. P. (1960). The use of advance organizers in the learning and retention of meaningful verbal
material. Journal of educational psychology, 51(5), 267.
Flanagan, B., & Ogata, H. (2018). Learning analytics platform in higher education in Japan. Knowledge
Management & E-Learning: An International Journal, 10(4), 469-484.
Hwang, G. J., Chu, H. C., & Yin, C. (2017). Objectives, methodologies and research issues of learning
analytics. Interactive Learning Environments, 25(2), 143-146.
Lan, A. S., & Baraniuk, R. G. (2016, June). A Contextual Bandits Framework for Personalized Learning
Action Selection. In EDM (pp. 424-429).
Li, M., Ogata, H., Hou, B., Uosaki, N., & Mouri, K. (2013). Context-aware and Personalization Method
in Ubiquitous Learning Log System. Journal of Educational Technology & Society, 16(3).
Okoshi, T., Tsubouchi, K., & Tokuda, H. (2019, July). Real-World Product Deployment of Adaptive Push
Notification Scheduling on Smartphones. In Proceedings of the 25th ACM SIGKDD
International Conference on Knowledge Discovery & Data Mining (pp. 2792-2800). ACM.
Pielot, M., Dingler, T., Pedro, J. S., & Oliver, N. (2015, September). When attention is not scarce-
detecting boredom from mobile phone usage. In Proceedings of the 2015 ACM international
joint conference on pervasive and ubiquitous computing (pp. 825-836). ACM.
Shimada, A., Okubo, F., Yin, C., & Ogata, H. (2017). Automatic Summarization of Lecture Slides for
Enhanced Student Preview Technical Report and User Study. IEEE Transactions on Learning
Technologies, 11(2), 165-178.
Truong, H. M. (2016). Integrating learning styles and adaptive e-learning system: Current developments,
problems and opportunities. Computers in human behavior, 55, 1185-1193.
Wan, S., & Niu, Z. (2018). An e-learning recommendation approach based on the self-organization of
learning resource. Knowledge-Based Systems, 160, 71-87.
Yin, C., & Hwang, G. J. (2018). Roles and strategies of learning analytics in the e-publication
era. Knowledge Management & E-Learning: An International Journal, 10(4), 455-468.
654
... where d(1) represents the prior probability of positive class data in the original data of the initial preschool language education data, d(0) represents the prior probability [14] of negative class data in the initial preschool language education data, and c represents the data balance factor. e effective classification process of preschool language education resource data is shown in Figure 2. ...
... Set the preschool education resource dataset to be integrated as M � m 1 , m 2 , . . . , m n , (14) where M represents the number of data from integrated preschool resources and m n represents the dimension of data elements. e above dataset to be integrated is divided into C class clusters, when the set class cluster membership matrix is represented as ...
Article
Full-text available
Aiming at the problems of high redundancy and slow integration speed in the existing education resource data integration methods, a new preschool language education resource integration method based on metadata warehouse is designed. The metadata warehouse is designed, and the advantages of the integrated database are analyzed. On this basis, the sample data of preschool language education resources are classified with the help of cost matrix, and the constraints of different types of classification are set. The data collector of preschool language education resources is set up by using random forest algorithm to complete the data collection of preschool language education resources. The data of preschool language education resources are processed consistently, and the convergence of the data is calculated by edge function. On this basis, the redundant data in preschool language education data resources are characterized with the help of discourse, and the redundant data are removed to complete the data preprocessing of preschool language education resources. We determine the dimension distance between preschool language education resource data and complete the clustering integration of preschool language education resource data with the help of fuzzy mean clustering algorithm. The experimental results show that the integration method designed in this paper can reduce the redundancy in the integrated data, and the integration speed is fast.
Conference Paper
Full-text available
The limited attentional resource of users is a bottleneck to delivery of push notifications in today's mobile and ubiquitous computing environments. Adaptive mobile notification scheduling, which detects opportune timings based on mobile sensing and machine learning, has been proposed as a way of alleviating this problem. However, it is still not clear if such adaptive notifications are effective in a large-scale product deployment with real-world situations and configurations, such as users' context changes, personalized content in notifications, and sudden external factors that users commonly experience (such as breaking news). In this paper, we construct a new interruptibility estimation and adaptive notification scheduling with redesigned technical components. From the deploy study of the system to the real product stack of Yahoo! JAPAN Android application and evaluation with 382,518 users for 28 days, we confirmed several significant results, including the maximum 60.7% increase in the users' click rate, 10 times more gain compared to the previous system, significantly better gain in the personalized notification content, and unexpectedly better performance in a situation with exceptional breaking news notifications. With these results, the proposed system has officially been deployed and enabled to all the users of Yahoo! JAPAN product environment where more than 10 million Android app users are enjoying its benefit. -- PDF at ACM DL : https://dl.acm.org/citation.cfm?id=3330732
Article
Full-text available
E-books have been introduced to educational institutions in many countries. The use of e-books in traditional classrooms enables the recording of learning logs. Recently, researchers have begun to carry out learning analytics on the learning logs of e-books. However, there has been limited attention devoted to understanding the types of learning strategies that students employ when they read e-books. In this paper, using e-book learning logs, we examine the learning strategies that students employed when reading e-books. In this paper, we will introduce how to identify learning strategies from e-book learning logs with two case studies. One is "Identifying Learning Strategies Using Clustering" and the other is "Examining Learning Strategies Using Sequential Analysis.". © 2018 Hong Kong Bao Long Accounting And Secretarial Limited. All Rights Reserved.
Article
Full-text available
In recent years, learning analytics has become a hot topic with many institutes deploying learning management systems and learning analytics tools. In this paper, we introduce learning analytics platforms that have been established in two top national Japanese universities. These initiatives are part of a broader research project into creating wide-reaching learning analytics frameworks. The aim of the project is to support education and learning through research into educational big data accumulated on these platforms. We also discuss the future direction of our research into learning analytics platforms. This includes introducing a model in which learning analytics tools and the results of research can be shared between different education institutes.
Conference Paper
Full-text available
Boredom is a common human emotion which may lead to an active search for stimulation. People often turn to their mobile phones to seek that stimulation. In this paper , we tackle the challenge of automatically inferring boredom from mobile phone usage. In a two-week in-the-wild study, we collected over 40,000,000 usage logs and 4398 boredom self-reports of 54 mobile phone users. We show that a user-independent machine-learning model of boredom –leveraging features related to recency of communication, usage intensity, time of day, and demographics– can infer boredom with an accuracy (AUCROC) of up to 82.9%. Results from a second field study with 16 participants suggest that people are more likely to engage with recommended content when they are bored, as inferred by our boredom-detection model. These findings enable boredom-triggered proactive recommender systems that attune their users' level of attention and need for stimulation.
Article
In e-learning, most content-based (CB) recommender systems provide recommendations depending on matching rules between learners and learning objects (LOs). Such learner-oriented approaches are limited when it comes to detecting learners’ changes, furthermore, the recommendations show low adaptability and diversity. In this study, in order to improve the adaptability and diversity of recommendations, we incorporate an LO-oriented recommendation mechanism to learner-oriented recommender systems, and propose an LO self-organization based recommendation approach (Self). LO self-organization means LO interacts with each other in a spontaneous and autonomous way. Such self-organization behavior is conducive to generating a stable LO structure through information propagation. The proposed approach works as follows: firstly, LOs are simulated as intelligent entities using the self-organization theory. LOs can receive information, transmit information, as well as move. Secondly, an environment perception module is designed. This module can capture and perceive learner's preference drifts by analyzing LOs’ self-organization behaviors. Finally, according to learners’ explicit requirements and implicit preference drifts, recommendations are generated through LOs’ self-organization behaviors. Based on applications to real-life learning processes, the ample experimental results demonstrate the high adaptability, diversity, and personalization of the recommendations.
Article
This paper is an extension of the work that was originally reported in [1]. In this paper, we propose a novel method of summarizing lecture slides to enhance preview efficiency and improve students’ understanding of the content. Students are often asked to prepare for a class by reading lecture materials. However, this does not always produce good results because the attention span of students is limited. We conducted a survey involving preview of lecture materials by more than 300 students and found that they want summarized materials to preview. Therefore, we developed an automatic summarization method to reduce the original preview materials to a summarized set. Our approach is based on the use of image processing and text processing to extract important pages from lecture materials, and then optimizing the selection of pages in accordance with a specified preview time. We applied the proposed summarization method to lecture slides. In our user study involving more than 300 students, we compared the relative effectiveness of the summarized slides and the original materials in terms of quiz scores, preview achievement ratio, and time spent previewing. We found that students who previewed the summarized slides achieved better scores on pre-lecture quizzes even though they spent less time previewing the material.
Article
Learning styles which refer to students’ preferred ways to learn can play an important role in adaptive e-learning systems. With the knowledge of different styles, the system can offer valuable advice and instructions to students and teachers to optimise students’ learning process. Moreover, e-leaning system which allows computerised and statistical algorithms opens the opportunity to overcome drawbacks of the traditional detection method that uses mainly questionnaire. These appealing reasons have led to a growing number of researches looking into the integration of learning styles and adaptive learning system. This paper, by reviewing 51 studies, delves deeply into different parts of the integration process. It captures a variety of aspects from learning styles theories selection in e-learning environment, online learning styles predictors, automatic learning styles classification to numerous learning styles applications. The results offer insights into different developments, achievements and open problems in the field. Based on these findings, the paper also provides discussion, recommendations and guidelines for future researches.
Article
It is hypothesized that learning and retention of unfamiliar but meaningful verbal material can be facilitated by the advance introduction of relevant subsuming concepts (organizers). 2 groups of 40 undergraduate Ss each were equated for sex, field of specialization, and ability to learn unfamiliar scientific material. Experimental Ss studied a 500-word-passage that supplied the subsuming concepts. Control Ss studied traditional type historical material of equal length. The learning material was then presented. Comparison of mean retention scores of the 2 groups unequivocably supported the hypothesis. From Psyc Abstracts 36:01:1CI67A. (PsycINFO Database Record (c) 2006 APA, all rights reserved).
A Contextual Bandits Framework for Personalized Learning Action Selection
  • A S Lan
  • R G Baraniuk
Lan, A. S., & Baraniuk, R. G. (2016, June). A Contextual Bandits Framework for Personalized Learning Action Selection. In EDM (pp. 424-429).