=Paper=
{{Paper
|id=Vol-2689/paper8
|storemode=property
|title=Context Aware Recommender Systems and Techniques in offering Smart Learning: A Survey and Future work
|pdfUrl=https://ceur-ws.org/Vol-2689/paper8.pdf
|volume=Vol-2689
|authors=Kevin Otieno Gogo,Lawrence Nderu,Stephen Makau Mutua
|dblpUrl=https://dblp.org/rec/conf/seia-ws/GogoNM20
}}
==Context Aware Recommender Systems and Techniques in offering Smart Learning: A Survey and Future work==
https://ceur-ws.org/Vol-2689/paper8.pdf
Context Aware Recommender Systems and
Techniques in offering Smart Learning: A Survey
and Future work
Kevin Otieno Gogo, Lawrence Nderu, Stephen Makau Mutua
Department of Computer Science School of Computing and Information Tech. School of Computing and Informatics
Chuka University JKUAT University MUST University
Chuka, Kenya Juja, Kenya Meru, Kenya
kevingogo2002@gmail.com lnderu@jkuat.ac.ke, smutua@must.ac.ke
Abstract— There exist several context aware collected data. The collected bits of information is what the
recommender systems (CARS) that have been designed system interprets into a meaningful action, based on the
to perform specific tasks in facilitating smart learning. current status of the entities that interact with the system.
Such context aware recommender systems have Context information that the system collects could refer and
potentially played important roles in education, not limited to identity, time, temperature, mood, location,
especially in recommending to a learner certain activity, environment, etc. In order to recommend relevant
activities. In this research we reviewed existing context learning content to learners in smart learning environment,
there is need to incorporate CARS to pick learners
aware recommender systems used in facilitating
preferences and likes.
pedagogical smart learning. Our survey paper addressed
Author [24], [13], [58], and [40], all clarify that currently
the following research questions; which contexts are used
learning activities have gone digital in most learning
in smart learning, how the contexts are collected, institutions, with most of the learners preferring to use smart
recommended activities, data mining techniques, and learning platforms. Thus calling for the need to
future work in CARS for smart learning. In our findings automatically recommend to a learner certain activities at
we addressed the research questions with reference to any particular time of learning. This would in turn reduce the
smart learning. We identified that there are numerous time taken to manually searching for the relevant learning
context aware recommender systems; but despite all activity. The recommended activities could range from
existing CARS in smart learning, there are several books for a certain level, courses to undertake, professionals
challenges and gaps that are still existing. Such in a certain field etc.
challenges include: - Absence of CARS that perfectly fits These activities are recommended based on the contexts
the ever changing learner’s needs and preferences, and collected from a learner. Such that to recommend
lack of standard database for smart learning, among professional to assist a learner, we may collect the location
other issues highlighted in our future work. details and level of expertise of a learner. Hence pointing out
that there are diverse contexts and context variables that can
Index Terms — Artificial intelligence, Context aware be used in context aware systems. There are several research
recommender system, Technology enabled Learning that have been done in CARS in smart learning, with each
research addressing a particular research problem. In this
I. INTRODUCTION paper we surveyed existing literature in CARS for smart
learning that were in line with our research questions.
A uthor [16], [25], and [40], explain that there is
massively huge educational content that has been
digitised in digital libraries and other learning platforms II. BACKGROUND
globally. On contrary there are insufficient mechanisms put In our background we gave an insight of smart learning,
in place to recommend relevant and personalized learning context aware recommender systems, environment and
content to learners. Specifically based on the learners’ ever techniques in smart learning. Context aware recommender
changing preferences, that we consider to be smart learning. system components, how smart learning contexts are
Reference [5], [45], and [9], point out that context aware collected from the environment and considerations when
systems collect a variety of information from their designing smart learning context aware recommender
environment and adapt their behavior according to the systems.
Copyright © 2020 for this paper by its authors. Use permitted under Creative Commons License Attribution 4.0 International
(CC BY 4.0).
A. Overview of smart learning based, demographic-based, utility-based, context-aware,
Reference [10], [47], [22], and [42], all point out that trust-aware, collaborative filtering, social network-based,
smart learning is student concentric strategy mode of and hybrid technique. The choice of a technique depends on
learning in provision of learning materials, where a learner the way contexts are to be collected and processed.
will have diverse choices of learning at his fingertip, and Such that for group based recommendation, the
learning can be achieved at home, field, school, etc. All these recommendation made are similar for members of a
depends on the learner’s choice of learning. particular group, while knowledge based will differ
Author [37], [52], and [8], explain that a smart learning depending on learner’s knowledge level. Demographic
system should be able to learn the users’ preferences based based recommendation will focus on some demographics,
on the environment of the learner. This is due to learners’ say how many times you have rated a certain item. While
ever changing demands, hence it is imperative to keep hybrid recommendation, combines more than one
collecting the learner’s choice of preferences every time. The recommendation technique.
collected learner preferences, can then be automated based C. How smart learning contexts are collected from the
on the way that data was collected and the expected results. environment
Source [9], [27], [44], and [57], suggest that technology
Papers [23], [41], [33], and [39], all point out that there are
enabled smart learning process usually takes the following
several ways that have been put across to collect contexts for
salient steps.
learning systems. The collection depends on the
1) Sensing environment and type of context to be collected. These
It entails scanning the user and his environment then contexts can be collected through: - Wi-Fi networks, RFID
collecting/sensing his particular requirements. (Radio frequency identification), wireless sensor networks
2) Reasoning (WSNs), internet, mobile agents, mobile and wearable
An array of artificial intelligence tools that analyze the sensors, and computer systems.
collected data using the sensors
3) Acting D. Considerations when designing smart learning context
It evolves around reacting to the environment, based on aware recommender models
the solutions offered by the reasoning stage Reference [28], [50], [51], [16], and [4], all point out that
when designing smart learning context aware recommender
B. Context aware recommender system components,
system, then the following factors are vital to consider for a
environment and techniques
more comprehensive and user-friendly recommender
Papers [34], [22], [17], and [5], all show that models of system. These considerations are diversity of
context awareness systems starts by first sensing their recommendations, recommender persistence, user privacy,
environment or contexts. During the sensing, they will user demographics, recommendations serendipity, and trust.
collect the contexts that they are adapted to collect. They
will then use their artificial intelligence to give an output, or III. LITERATURE REVIEW
adapt to the environment in a manner that is proportional to
the collected contexts. Thus different contexts collected will Due to the evolution of MOOC (Multiple Open Online
necessitate the context aware model to behave in a particular Courses) available in the internet, with each course having
fashion. Likewise [26], [49], and [3], suggest that there are a its own objective. There exist several CARS in smart
number of environment awareness that context awareness learning systems, each addressing a pertinent goal, justifying
systems could perceive. These context aware environments the complexity and the several activities that can be
could be grouped as: - recommended in smart learning. Hence we restricted our
survey to CARS papers that addressed the following research
1) Physical environment context questions in pedagogical smart learning: -
This is where the context awareness data is one that is
related to physical world parameters. Such as location of Which contexts are used in smart learning,
user, body temperature of user etc. How are the contexts collected,
2) Human environment context (user context) What data mining and recommendation
These contexts are pertaining to the user; such as: -User's techniques used to processes the contexts,
identity, habits of the user, to-do-list of the user etc. What are the recommended activities,
3) ICT context or virtual environment context. Future work in CARS.
This is where the context awareness data is data relating
to a device in the distributed system. Such as GPS During our literature search we identified several
location etc. articles on context aware recommender systems in learning.
A. Selection of articles for review
Reference [60], [33], [12], research suggest that there are
As outlined by [4], and [45], there are several research that
several recommendation techniques that have been proposed
have been carried out in CARS smart learning; with each
and used by a number of research. These recommendation
recommending an activity and addressing their own gap.
techniques include group-based, knowledge-based, content-
We searched through various journals, and conference 3) Ubiquitous computing technologies and context aware
proceedings using the following terminologies: - recommender systems for ubiquitous learning
Context aware recommender systems in learning, Papers [35], and [46], Developed ubiquitous learning
Recommender systems in smart learning system. The system integrates context aware and ubiquitous
computing technologies. Their system, a 2D or a QR bar
Context aware in education
code that enabled students in getting recommendations on
Several journal articles and conference proceedings were the best herb that can be used for a particular ailment. The
availed. We selected journals published within this decade various plants have an attached QR code, from which the
and specifically addressing majority of our research students can scan the plants QR code with their mobile
questions. devices to identify the herb and to an extent the medicinal
B. Surveyed papers on context aware recommender values of the herbs. They used clustering algorithm and
systems for smart learning association rules to unearth relationships in databases.
Collaborative filtering approach assists the algorithm to
Below are some of the papers that we reviewed in smart
identify the relationships.
learning contexts aware recommender systems. Where we
had more than one author with similar research but different
4) Keywords Learning Materials Recommendation
findings, we documented all their findings under one title.
Framework
Reference [21], and [59], explain that there are a number of
1) Context-Aware Recommender Systems for E-Learning:
E-learning material recommendation frameworks. In their
A Survey and Future Challenges
framework an instructor uploads learning content in the
Author [15], [32], and [53], all argue that several research
database. Each content was given a keyword which will be
have been conducted, on technology enabled learning
used for recommendation purposes. The keyword will have
recommender systems. Majority of the studied models use
content attributes such as title, author etc. A student will be
knowledge-based filtering, collaborative filtering and
given recommendations based on the keyword that the
hybrid techniques. These techniques depend on
student has requested. There was a learner’s ratings for
recommendations that are to be suggested to a learner.
these keywords. Further recommendations are given based
There are diverse contexts that are usable in smart learning.
on average learner ratings of these keywords. Otherwise,
These contexts are categorized as physical condition
the recommender system will predict a good learners’
contexts, computing contexts, user contexts, resources
ratings for the item.
contexts, time contexts, situation contexts etc.
Several activities were recommended with these
5) Book Recommender System using Fuzzy Linguistic
recommender systems. These recommender systems were
Quantifier and Opinion mining
to suggest any of the following activities:- teachers, peers,
Author [2], [14], and [57], proposed a book recommender
learning content, knowledge levels, training needs, books,
system where learners give opinion and rate books, thus a
institutions, lighting levels, learning styles, courses to
customer review for the available books. The customer will
undertake etc.
give their opinion review online. The reviews will be
Their research identified some challenges and future
aggregated, and an ordered weighted average (OWA) will
work, which include: - combining more than one context
be determined. A book with the highest OWA is usually
groups to have a hybrid, contexts that have not yet been
the one with the highest recommendation. This context
used, automatic capturing of learners’ contexts, user
aware recommender suggests a number of books, and let a
interests, and standard dataset for smart learning
user rate the books. It then recommends books with similar
recommendations. They also suggested other data mining
OWA ratings as those that the learner rated highly in his
methods other than the mostly used nearest neighbour,
ratings. This is not practical where there are several books.
crowdsorcing, pattern matching and association rules.
6) Fuzzy Logic Based Context Aware Recommender for
2) Automatic learner needs identification using context
Smart E -learning Content Delivery
awareness and artificial intelligence
Reference [22], and [39], suggest that Fuzzy logic-based
Reference [9], [29] and [56] explain that context awareness
context aware recommender for smart E-learning content
could be used to investigate learners’ behavior pattern and
recommender, is a model that collect learner’s context and
offer them relevant solutions with regard to their learning
give suggestions of relevant learning material to a learner.
behavior. Their systems uses artificial intelligence and
A learner will be provided with an assessment test. The time
context awareness to provide learners with their specific
taken to complete the test and the assessment test score, will
likes. These likes, depends on their past data and the data
form the learner’s contexts.
collected from their ubiquitous devices. However, these
The collected contexts will be analyzed using type 1 fuzzy
systems are limited to recommending to a learner, the
logic, to output a learner’s knowledge level. The output
reading content based on their previous likes unfortunately
knowledge level from the fuzzy logic will be assigned to a
their systems do not support dynamic learner preferences.
learner, which will then be used to recommend relevant
learning content for the learner. There will be a database of a group of users and recommend to a user with reference
with learning contents OWA scored with the professionals to a group that a user best fits.
of that learning subject. Thus learning content with OWA
of + 10 from a learner’s knowledge level will be 10) Enhanced Recommendations for e-Learning Authoring
recommended to the learner. Tools based on a Proactive Context-aware Recommender
Author [19], [20], and [30] suggest that authoring tools are
7) Dynamic context aware learning objects in e-learning powerful systems in the area of e-Learning. They assist
environment teachers to create new learning objects by reusing or editing
Reference [38], [2], and [43], elaborate that in context existing educational resources that are stored in learning
aware E-learning environment, we could have either a repositories. However, due to large number of resources
random user or an adequate user of E-learning system. We these tools access, it is usually difficult for teachers to find
have three modules that support the interpretation of user the most suitable resources for their target level.
contexts for all the users. These are the system contexts, The papers propose models that could generate proactive
user contexts, and environment contexts. context-aware recommendations on resources during the
The existing context aware systems for E-Learning have creation process of new learning objects. The models use
no standardized set of context parameters, hence the context topics of the learning object and the target audience.
selection of contexts is based on randomly considered set The models would then suggest suitable resources without
of parameters. Some of the context parameters include: - explicit user request being needed. Therefore, the user
Learner’s personal profile, level of expertise, preferences, would discover the resources during the creation process.
intention, quality of learning service, learner’s devices,
pace, state etc. That are used for recommendations. 11) A Fuzzy Approach to Multidimensional Context Aware
They suggest that there is need to have a standard model e-Learning Recommender System (CA-ELRS)
incorporating all the three modules, in order to integrate all Reference [18], and [55], point out that majority of the
the eLearning learning objects to achieve dynamic contexts. learners context include: - mood, emotion, duration of
study, place, time, and social interaction. The paper thereby
8) SMS based E-Assessments enabling better Student works towards developing an effective multidimensional
Engagement, Evaluation and Recommendation Services in CA-ELRS using mood and time duration as fuzzy logic
E-Learning using Fuzzy Rules and Course Ontologies input concepts. Then through use of fuzzy logic, the crisp
Author [6], and [36], research proposed models where output can be given, which will then be used to match the
students interact with teachers as usual in class or using e- relevant learning content.
learning platforms. The models suggest that students will
answer quizzes pertaining to their interactions with the 12) Context-aware recommender for mobile learners
teacher. The way a student answers the quizzes will now Author [10], and [54], research entail building mobile
guide the student on whether to proceed or revisit the earlier recommender systems with semantic-rich awareness
topics. The model mines a student SMS response on a fuzzy information. Their content recommendations are tailored to
logic, which then gives the student personal a learner’s background, context, and task at hand of the
recommendation whether to proceed to the next topic or mobile learner. These content recommendations are also
not. The research is limited to course topics, and whether integrated with different operational characteristics such as
the student is competent to proceed to the next topic based varying network bandwidth and limited mobile devices
on the SMS E-assessment score. resources. In incorporating all the above factors in one way
or the other, it was therefore necessary to consider a
9) E-learning group recommender systems: Combining proactive context awareness mechanism that can sense both
User-User and Item-Item Collaborative Filtering system-centric and learner-centric context and adapt the
Techniques accessed services on the go. The models use shared
References [48], [1], and [11], proposed that a group ontology space and unified reasoning mechanism.
recommender system has combined a hybrid of item-item
and user-user collaborative filtering techniques to build 13) Constructing a User-Friendly and Smart Ubiquitous
their recommender systems. Implying that instead of Personalized Learning Environment by Using a Context-
recommending based on similar users only, they also Aware Mechanism
incorporate both similar users and similar items in their Author [31] and [7] developed intelligent personalized
group collaborative filtering recommendations. They context-aware recommendation learning systems. The
further argue that majority of recommender systems were systems comprised of multimedia streaming learning
group recommenders and not personal recommenders. subsystem, and context-aware learning subsystem. The
Noting that it would be difficult to build personal learning content being recommended change with the
recommenders for each user. Their have not built context learner’s location and environment change. They provided
aware recommender system for every individual, but have dynamically adjusted learning that is location-based. The
built a recommender system that consider the preferences study used broadly defined term personalized mobile
learning due to various sensor technologies that is usually IV. RESULTS AND DISCUSSIONS
incorporated in mobile equipment. These recommendations
A. Results
were personalized learning, that change according to the
learner’s environmental information such as history, From the above literature review, we managed to tabulate
regional traits, nearby buildings etc. the result in table 1 below for easy visualization and quick
follow-up, based on each CARS studied in smart learning.
TABLE 1
A SUMMARY OF LEARNER CONTEXTS, RECOMMENDER SYSTEMS, DATA MINING TECHNIQUES, RECOMMENDED
ACTIVITIES AND FUTURE RECOMMENDATIONS USED IN THE DEVELOPMENT OF SMART LEARNING CARS
Main Contexts How contexts Recommender Recommendation Recommended Challenges and
research collected would be system (s) / data mining Activities Future work
Publication(s) collected Technique(s)
[32] Verbert, - Software - Computers - Location based - Nearest - Suggest teachers, - Use of hybrid
K. et al. - Task - GPS learning. neighbour, peers, knowledge recommendation
and - Proximity - Mobile phones - System based - Crowdsorcing, levels, techniques
[15] Chughtai - Knowledge - Sensors learning. - Pattern matching - Lighting styles - Standard dataset
M.W, et al. - Learning styles - Mobile agents - Time based etc. - Association rules - Learning content, - Automatic context
– Interest etc. levels etc capture
[9] Aztiria, A., - Past likes & - Timers - Keywords - Pattern matching - Learning content - Incorporate several
et al. behavior data - Computer Learning based on learners behaviors
- Current - Social media Materials likes and behavior - Cold start, especially
behavior data - RFID when few past learner
behavior are available
[35] Thiprak, - QR barcode - Bar code -UbiCARsUL - Clustering - Identify a herb - Integration with other
S., and - 2D bar code reader algorithm - Identify extent of capturing devices
Kurutach, W. - Computer medicinal values in
- Internet herbs
[21] Ghauth, -Learner ratings - Computer - learner needs - Context - Reading content - Automatic context
K. I., and - Keywords - Search identification awareness - Reading content capture
Abdullah, N. engines - Normal search ratings
A.
[2] Adnan, - Learner’s book - Computer - Book - Aggregated - Suggests reading - Capturing several and
M., N., et al. opinion - books in recommender ordered weighted book diverse opinions -
- Book ratings database ratings system average on a book Challenges during
information overload
[22] Gogo, K. -Learner test - Learner -Smart e-learning - Context aware - Relevant and -Use of interval type 2
O., Nderu, L. assessment score assessment content - Fuzzy logics personalized learning fuzzy logic
and Mwangi, -Time to - Computer recommender content - Incorporate other
R. W. complete - Timer recommendation learner contexts
assessment test - Learning information
overload
-Network delays
[38] Zarrad -User context, - GPS - Dynamic -Unspecified - Relevant e-learning -Need for a standard
A., and Environment - Computer learning objects content model that would be
Zaguia, A. context, - Smart phone in e-learning able to integrate all the
-System context - Wireless eLearning learning
networks objects
- Wearable - Standard location
sensors based database.
[6] Antony, J., - SMS E- - Learner -SMS based E- -Fuzzy logic - Suggest learning - Research is only
et al. assessment score assessment Assessments topic to visit limited to course
- Teacher - Computer Recommendation topics. It needs to be
interaction - Smart phone Services in E- expanded to capture
Learning other learning items
[48] Pujahari, - User - Preference -E-learning group -Hybrid of similar - E learning content -Build personal
A., preferences on ratings on recommender users and similar based on similar recommenders for each
and Padmanab item learning content based on User- items users and similar user
han, V. - Computer User and Item- - Collaborative items
Item filtering
[19] and [20] - User context - Mobile device -Proactive -Aggregating - Proactive relevant - Need to integrate
Gallego, D., et (idle, eyes open, - Camera Context aware context scores for authoring tools for e- several real world
al. reading, etc.) - Computer Recommender each collected learning learning contexts
-Learning object - learning for e-Learning context, then
context (level, content access Authoring Tools compare with the
topic language,) -Internet most closest score
[18] Dwivedi -Learner moods - Mobile device Multidimensional - Item based - Suggest learning - Incorporation of
P., and (happy, sad, - Camera Context Aware e- collaborative content with regard several moods and
Bharadwaj moody, etc.) - Computer Learning filtering to learners moods emotions contexts
K.K. Recommender
[10] Benlamri, -Mobile learner - GPS Context-aware - Ontology -Recommend - Efficient context
R., and Zhang, surrounding - Computer recommender for reasoning learning content to integration and
X. contexts (GPS, - Google maps mobile learners mobile learners, adaptation
wireless nets, - Smart phone depending on mobile - Sufficient context
security, etc.) surrounding context change adaptation
[31] Yao, -GPS location - Barcode -Personalized -Intelligent -Location based M- - Dynamic location
C.B., -QR code - GPS context-aware personalized learning, for changes
- Computer recommendation context-aware recommending - Mapping several
- - Wi-Fi (PCAR) learning learning algorithm learning content locations.
- Wearable system - Aggregations of based on location
sensors contexts
algorithm, normal search, ordered weighted average and
B. Discussions
collaborative filtering.
As explained by Author [18], [40], [54], [46], and [55], as Our results in table 1 further indicate that each and every
well as from the results in table 1 of our study. We identified context aware recommender system in learning is addressing
that smart learning incorporates e-learning, mobile learning, a particular goal, thus diverse future work with regard to their
pervasive learning as well as ubiquitous learning. Each of the individual goals. Some of the items that have been
above learning has its own set of tools and architectures. With recommended in smart learning include: - recommendations
the evolution of MOOC and the various learning styles and of learning material, teachers, peers, topics, learning content,
learner preferences. There exists several contexts that can be lighting styles, levels of learning, etc. In our survey, we noted
used in designing context aware recommender systems, that there exist several recommender systems that recommend
depending on the contexts that can be collected, as well as the similar activities, say learning content; but with different
learning architecture to be developed. contexts collected from a learner. This then confirms that, we
These learning contexts can be grouped as user context do not need to collect same contexts for similar
(preference, intention, moods, knowledge level, expertise, recommendations.
frequency, goals, keywords, language, background etc.). Similarly, same contexts could be collected from learners
Environment context (GPS location, region, temperature, using different CARS, but lead to different activities being
time, proximity, day etc.), or System context (hardware, recommended to a learner. For example, in order to
software, network, system speed, etc.). recommend relevant books to a particular learner, any or a
Due to the various contexts in CARS in smart learning, combination of the following contexts can be collected from
there are diverse strategies of collecting user contexts in smart learners: - learner preference, keywords, level of study,
learning. These strategies depend mainly on the CARS expertise, etc. The contexts collected depend entirely on the
architecture and environment from which the contexts are to architecture, data mining method, user profiles, context
be collected. Such collection methods include: - the use of variables, and strategies for collecting learners’ contexts.
computers, Wi-Fi, wearable sensors, QR code, internet, There are still several areas of research that have not been
RFID, touch screens, etc. User contexts tends to use majorly fully exhausted regarding CARS in smart learning, especially
wearable devices and computers. Environment contexts tend with the evolution of MOOC and learner centric learning. For
to use gadgets that can collect environmental contexts, such example, in a research to recommend relevant books to a
as GPS, Wi-Fi, thermometers etc. As system contexts tend to learner, when one has picked similar contexts. Then the
use computers, timers, internet etc. following could be issues of concern.
We also noted that there are various recommendation and Taking:-
data mining techniques, which extracts useful information Strategies for collecting learners’ contexts …….…….A
from learner contexts and guide the recommendation Recommendation techniques and data mining techniques
intelligence. These recommendation techniques used include used in CARS ……………..…………….……..……. B
content-based, group-based, knowledge-based, demographic-
based, utility-based, trust-aware based, social network-based, If there are 7 strategies through which specific learner
fuzzy-based, collaborative filtering, and hybrid contexts can be collected, and 6 data mining techniques that
recommendation. On the other hand, the data mining can be used for the research. So A=7, and C=6. Then this
techniques that can be used in mining learning content in research alone would have 7x6 = 42 possibilities. If we decide
smart learning include: - nearest neighbour, regression, to diversify the contexts, then even more possibilities will be
crowdsorcing, pattern matching, association rules, clustering available exponentially for the same research.
Following the fact that there are several parameters to be REFERENCES
addressed in CARS in smart learning, then there arises several [1] Abel, F., Bittencourt, I.I., Costa, E., Henze, N., Krause, D., and
possibilities, challenges, and future work. Vassileva, J., “Recommendations in Online Discussion Forums for E-
Learning Systems,” IEEE Trans. Learning Technologies, vol. 3, no. 2,
pp. 165-176, Apr.-June 2010.
V. CONCLUSION AND RECOMMENDATIONS [2] Adnan, M., N., et al, Content Based News Recommendation System
Based on Fuzzy Logic, 3rd International Conference on Informatics,
A. Conclusion Electronics & Vision, IEEE (2014) DOI 978-1-4799-5180-2/14
This survey justifies that there are several context aware [3] Adomavicius, G., and Tuzhilin, A., (2011), Context-Aware
recommender systems in smart learning, with several areas Recommender Systems. In Recommender Systems Handbook, pages
217-253, Springer, 2011
that have not been fully researched. This can be attributed [4] Aggarwal, C. C., Recommender systems, the textbook, Springer
by the several learner contexts available; various International publishing Switzerland, 2016.
recommendation and data mining techniques, various [5] Angham, A., et al., GECAF: a framework for developing context-
strategies through which contexts can be collected, and the aware pervasive Systems, Springer-Verlag Berlin, Heidelberg
Computing Sci Res Dev, 2013, 30:87–103
numerous activities that can be suggested in smart learning. [6] Antony, J., et al, SMS based E-Assessments enabling better Student
This study further points out that due to the evolution of Engagement, Evaluation and Recommendation Services in E-
MOOC and learner centric learning, the aspect of CARS in Learning making use of Fuzzy Rules and Course Ontologies, 4th
offering smart learning is even being complicated with International Conference on Technology for Education, IEEE (2012)
pp 119-125
technological advancements. These are due to the random [7] Asabere, N. Y., Towards a Viewpoint of Context-Aware
set of parameters, several existing context variables that a Recommender Systems (CARS) and Services, International Journal of
CARS can choose, and diverse learner preferences. Hence Computer Science and Telecommunications, vol. 4, pp. 19-29, 2013
there is no set standard for building CARS in smart [8] Ashley-Dejo, E., Ngwira, S., and Zuva, T., A survey of Context-aware
Recommender System and services in Computing, Communication
learning, but open to individual researcher. and Security (ICCCS), IEEE Pamplemousses International
The survey further illustrate that there exist numerous Conference on 4-5 Dec. 2015, pp 1 – 6.
contexts, context variables, data mining techniques, [9] Aztiria, A., et al. Learning patterns in ambient intelligence
strategies for collecting contexts and CARS for smart environments: a survey, Springer Science and Business Media B.V.,
Artificial Intelligence Rev (2010) pp 34:35–51
learning. But despite the various context aware [10] Benlamri, R., and Zhang, X., Context-aware recommender for mobile
recommender systems in place, there is no recommender learners, Springer open journal (2014), Human-centric Computing
system that fully fits every learner’s preference. This is due and Information Sciences 2014, 4:12
to the ever changing learner preferences, contexts and [11] Yong, J., Lee, H.Y., H. Yoo, K., Youn, H.Y., and Song, O., (2011)
Personalized recommendation system reflecting user preference with
learning environments. context-awareness for mobile TV, in Proc. 9th IEEE Int. Symp.
Parallel Distrib. Process. Appl. Workshops, 2011, pp. 232_237.
B. Recommendations and future work
[12] Cantador, I., and Castells, P., Semantic contextualization in a news
There exists different CARS research in smart learning as recommender system. In Workshop on Context-Aware Recommender
indicated in table 1, although each context aware Systems (CARS 2009). New York, 2009.
[13] Chen, C., et al. Real-time Smartphone sensing and recommendations
recommender system is addressing a specific goal. Even towards context-awareness shopping, Springer-Verlag Berlin
though the CARS could have same goals, they would have Heidelberg, Multimedia Systems, 2013, 21: 61–72
utilized different set of contexts, diverse recommendation [14] Chen, G., D., and Chao, P.Y., Augmenting Traditional Books with
techniques, or different strategies through which contexts are Context-Aware Learning Supports from Online Learning
Communities, Educational Technology and Soc., vol. 11, no. 2, pp.
collected. 27-40, 2008.
Due to the dynamic and ever changing learners’ [15] Chughtai M.W, et al, E-learning recommender systems based on goal-
preferences and contexts, the following gaps arise: - based hybrid filtering. Int Journal of Distrib Sens Netw (2014):1–19
Automatic extraction of learner contexts especially for [16] Dagger, D., Conlan, O., and Wade, V., Fundamental Requirements of
Personalised Elearning Development Environments, Proc. World
big databases without cold start problems, Conf. E-Learning in Corporate, Govt., Healthcare, and Higher
Capturing ever changing learners’ contexts in real time, Education, pp. 2746-2754, Oct. 2005.
Development of a situation aware CARS that can be [17] Dengpan Y. et al, Mobile crowd-sensing context aware based fine-
grained access control mode, multimedia tools application. Springer,
able to perceive every context situation (i.e. if the Multimed Tools Appl 2010, 1: 1-17
environment is noisy / unconducive for learning hence [18] Dwivedi P., and Bharadwaj K.K. (2013) A Fuzzy Approach to
postpone learning), Multidimensional Context Aware e-Learning Recommender System.
In: Prasath R., Kathirvalavakumar T. (eds) Mining Intelligence and
Mapping of several locations in context aware data, Knowledge Exploration. Lecture Notes in Computer Science, vol
Sharing of recommendations or contextual information 8284. pp 600-610, 2013. © Springer publishing Switzerland
[19] Gallego, D., Barra, E., Gordillo, A. and Huecas, G.; Enhanced
over big data databases, recommendations for e-Learning authoring tools based on a proactive
Standard context capturing devices in smart learning, context-aware recommender, 2013 IEEE Frontiers in Education
Developing a proactive CARS other than the traditional Conference (FIE), Oklahoma City, OK, (2013), pp. 1393-1395.
[20] Gallego, D., Barra, E., Aguirre, S., and Huecas, G., A model for
user requests and response given scenarios. generating proactive context-aware recommendations in e-learning
From the above gaps, we recommend building of a smart systems, in Frontiers in Education Conference, 2012, pp. 1364–1369.
learning CARS that would be able to have a universal [21] Ghauth, K. I., and Abdullah, N. A., The Effect of Incorporating Good
Learners' Ratings in e-Learning Content based Recommender System.
framework / model and platform for smart learning, with a JSTOR, Educational Technology & Society, 2011, 14 (2), 248–257
standard database for contextual data and recommended data [22] Gogo, K. O., Nderu, L. and Mwangi, R. W., Fuzzy Logic Based
for every learner in smart learning. Context Aware Recommender for Smart E-learning Content
Delivery, 2018 5th International Conference on Soft Computing & [44] Mokbel, M.F., and Levandoski, J.J. Toward context and preference-
Machine Intelligence (ISCMI), Nairobi, Kenya, (2018), pp. 114-118. aware location-based services. In Proceedings of the Eighth ACM
[23] Gogo, K., and Opiyo, T. E., Mobile agent-based system for listing International Workshop on Data Engineering for Wireless and Mobile
fundamental Wi-Fi peer-to-peer network details, International Journal Access, pages 25–32. ACM, 2009.
of computer applications, Volume 114- No. 5, pp 44-50, March 2015, [45] Newby, P., Research Methods for Education, second edition, (2014)
ISBN 0975 – 8887 Routledge, London https://doi.org/10.4324/9781315758763
[24] Hervás, B., J., et al. Towards the Everyday Computing in the [46] Zhiwen Y., et al. Towards a semantic infrastructure for context-aware
Classroom through RFID in Computers and Education, E-Learning, e-learning, multimedia tools application. Springer, Multimed Tools
From Theory to Practice, Springer 2007, 143-154 Appl, 2010, 47:71–86
[25] Hossain, M. et al, Developing and validating a model explaining the [47] Uskov, V., et al, Smart Education and Smart E-Learning, Springer
assimilation process of RFID: An empirical study, Springer, 2014. Computers and Education, E-Learning, From Theory to Practice,
[26] Huang, Y.M., Huang, S.H., Lin, and Y.T.: A ubiquitous English Springer 2015, 14-24
vocabulary learning system: evidence of active/passive attitudes vs. [48] Pujahari, A., and Padmanabhan, V., Group Recommender Systems:
usefulness/ease-of-use.Comput. Educ. (2012) 58(1), 273–282 Combining User-User and Item-Item Collaborative Filtering
[27] Hussein, T., Linder, T., Gaulke,W., and Ziegler, J.,(2009), Context- Techniques, International Conference on Information Technology
aware recommendations on rails. In Workshop on Context-Aware (ICIT), IEEE at Bhubaneswar on 21-23 Dec. 2015, pp 148 – 152,
Recommender Systems (CARS 2009), New York, 2009. [49] Punam, B., and Agarwal, S., A situation- aware proactive
[28] Jannach, D., Zanker, M. and Friedrich, G., Tutorial: Recommender recommender system, 12th International Conference on Hybrid
Systems. International Joint Conference on Artificial Intelligence Intelligent Systems, IEEE (2012) pp 85-89
Beijing, August 4, 2013 [50] Riboni, D. and Bettini, C., COSAR: hybrid reasoning for context-
[29] Jara, J. A. et al, Mobile discovery: discovering and interacting with the aware activity recognition, Springer Verlag London,
world through the Internet of things, PersUbiquitComput, 2014, PersUbiquitComput, 2011 15:271–289
18:323–338 [51] Ricci, F., Rokach, R., Shapira, B., and Kantor, P., B., (Editors),
[30] Jeremic, Z., Jovanovic, J., and Gasevic, D., Semantic Web Recommender Systems Handbook (2010) Springer
Technologies for the Integration of Learning Tools and Context- [52] Rutishauser U, Joller J, Douglas R., Control and learning of ambience
Aware Educational Services, pp. 860-875. Springer, 2009. by an intelligent building. In: IEEE on systems, man and cybernetics:
[31] Yao, C.B., (2017), Constructing a User-Friendly and Smart a special issue on ambient intelligence, IEEE systems, man, and
Ubiquitous Personalized Learning Environment by Using a Context- cybernetics society. 2005, pp 121–132
Aware Mechanism, IEEE transactions on learning technologies, vol. [53] Salman, Y., Abu-Issa, A., Tumar, I., and Hassouneh, Y., A Proactive
10, no. 1, January-March 2017 pp 104 -114 Multi-Type Context-Aware Recommender System in the Environment
[32] Verbert, K. et al, Context-Aware Recommender Systems for Learning: of Internet of Things; in Computer and Information Technology;
A Survey and Future Challenges, IEEE transactions on learning Ubiquitous Computing and Communications; Pervasive Intelligence
technologies, (2012) VOL. 5, NO. 4 pp. 318-335 and Computing, 2015 Liverpool IEEE International Conference on
[33] Yadav D. K. and Katarya, R., Study on Recommender System using 26-28 Oct. 2015 pp 351 – 355,
Fuzzy Logic, 2018 Second International Conference on Computing [54] Sappelli, M., Verberne, S., and Kraaij, W., Evaluation of Context-
Methodologies and Communication (ICCMC), Erode, (2018), pp. 50- Aware Recommendation Systems for Information Re-Finding, journal
54. DOI: 10.1109/ICCMC.2018.8487669 of the association for information science and technology, 68(4): 895–
[34] Yera, R., and Mart´ınez, L., A recommendation approach for 910, (2017) Wiley Online Library
programming online judges supported by data preprocessing [55] Shen, L., Wang, M., and Shen, R., “Affective E-Learning: Using
techniques, ApplIntell, Springer Science+Business Media NY (2016) Emotional Data to Improve Learning in Pervasive Learning
[35] Thiprak, S., and Kurutach, W., Ubiquitous computing technologies Environment,” Educational Technology and Soc., vol. 12, no. 2, pp.
and context aware recommender systems for ubiquitous learning, 176-189, 2009
Electrical Engineering, Computer, Telecommunications and [56] Zhu, H., et al; Mining Personal Context-Aware Preferences for
Information Technology (ECTI-CON), 2015 IEEE 12th International Mobile Users, 2012 IEEE 12th International Conference on Data
Conference, pp 1 – 6. Mining, pp: 1212 – 1217, (2013)
[36] Khodke, P. A. et al, Neuro Fuzzy Intelligent e-Learning Systems, 2016 [57] Sohail, S., S., Siddiqui, J., and Ali, R., Book Recommender System
Online International Conference on Green Engineering and using Fuzzy Linguistic Quantifier and Opinion Mining, J.M.
Technologies (IC-GET), IEEE (2016), pp 231-239 Corchado Rodriguez et al. (eds.), Intelligent Systems Technologies
[37] Klašnja-Mili´cevi´c A, Ivanovi´c M, Nanopoulos A., Recommender and Applications 2016, Springer 2016, Advances in Intelligent
systems in e-learning environments: a survey of the state-of-the-art Systems and Computing 530, pp 573-582
and possible extensions. Springer (2015), ArtifIntell Rev 44(4):571– [58] Soylu, A., et al. Context and Adaptivity in Pervasive Computing
604 Environments: Links with Software Engineering and Ontological
[38] Zarrad A., and Zaguia, A., Building a Dynamic Context Aware Engineering. Journal of Software, Vol. 4, No. 9, p 992-1013,
Approach for Smart e-Learning System, 2015 Second International November 2009
Conference on Computing Technology and Information Management [59] Sundermann, C. V., Domingues, M. A., Marcacini, R. M., and
(ICCTIM), IEEE pp144 – 149, 2015, ISBN: 978-1-4799-6211-2 Rezende, S., Using Topic Hierarchies with Privileged Information to
[39] Luo, J., Dong, F., Cao, J., and Song, A., “A Context-Aware Improve Context-Aware Recommender Systems, Intelligent Systems
Personalized Resource Recommendation for Pervasive Learning,” (BRACIS) Brazilian Conference on 18-22 Oct. 2014, pp 61 – 66, Sao
Cluster Computing, vol. 13, pp. 213-239, June 2010. Paulo DOI: 10.1109/BRACIS.2014.22
[40] Manouselis, N., Drachsler, H., Vuorikari, R., Hummel, H. and Koper, [60] Tarus, J. K., et al, Knowledge-based recommendation: a review of
R., “Recommender systems in technology enhanced learning,” in ontology-based recommender systems for e-learning, Springer
Recommender Systems Handbook, F. Ricci, L. Rokach, B. Shapira, Science -Business Media Dordrecht (2017), ArtifIntell Rev
and P. B. Kantor, Eds. Springer US, 2011, pp. 387–415.
[41] Mettouris, C., Achilleas P. A., and Papadopoulos, G. A., A Context
Modelling System and Learning Tool for Context-Aware
Recommender Systems, D. Hernández-Leo et al. (Eds.): EC-TEL
2013, LNCS 8095, pp. 619–620, 2013. Springer-Verlag
[42] Meyer, A., & Rose, D.H., The future in the margins: The role of
technology and disability in educational reform. In D. Rose, A. Meyer
and C. Hitchcock (Eds.). The universally designed classroom:
Accessible curriculum and digital technologies, Cambridge, MA:
Harvard Education Press, 2005, pp.13–35.
[43] Minu. M. et. al., Context Aware E-Learning System with Dynamically
Composable Learning Objects / (IJCSE) International Journal on
Computer Science and Engineering, Vol. 02, No. 04, 2010, 1245-1253