Personalized recommendation is important for online students to select rich learning resources and make their own learning schedules. We propose SARLR, a new self-adaptive recommendation algorithm of online learning resources. The SARLR algorithm integrates an IRT-based learning cognitive model named T-BMIRT into the recommendation framework and is able to adaptively adjust learning path recommendations based on dynamic of individual learning process. The experimental results show that the SARLR algorithm outperforms the existing recommendation algorithms.
With the growing prevalence of online education, students have access to all kinds of
electronic learning resources, including electronic books, exercises and learning videos.
Given the diversity of students’ background, learning styles and knowledge levels, it is
essential to have personalized recommendation tools to facilitate students in choosing
their own learning paths to satisfy their individual needs [
Most Intelligent Tutor Systems (ITS) such as [
Although the data-driven recommendation approach is more scalable and general than the rule-based approach, current proposed solutions have common problems in achieving highly adaptive recommendation towards students’ latent learning state. They often focus on either searching for similar learning resources based on content or identifying similar student groups based on their learning behaviors. The recommended learning objects or paths fail to consider the impact of difficulty of learning objects and dynamic change in students’ learning states.
In this paper, we propose a novel learning recommendation algorithm named SARLR, which attempts to integrate an IRT-based learning cognitive model into the recommendation framework and to adaptively adjust learning path recommendations based on dynamics of individual learning process. Specifically, we introduce a temporal, multidimensional IRT-based model named as T-BMIRT, which can accurately infer student proficiency of multiple latent skills and difficulties of exercise assessments. In addition, the T-BMIRT model incorporates the parameter of video learning, which can describe the improvement in student skills after their interactions with video lectures. Based on the T-BMIRT model, the SARLR algorithm can comprehensively analyze every student’s skill progress at each learning step and recommend to them a personalized learning path with the matching online video lectures and homework problems.
The contributions of this paper are the two-fold. First, we introduce the T-BMIRT model, to estimate students’ latent skill levels and difficulties of learning resources for recommendation. Second, we propose the SARLR algorithm by integrating the TBMIRT model in the adaptive recommendation process of learning resources. The experimental results confirm that the SARLR outperforms regular recommendation algorithms. Lastly, we present an evaluation strategy for recommendation algorithms in terms of rationality and effectiveness. 2
Data-driven learning recommendation algorithms often utilize common
recommendation methods widely adopted in the e-Commence area, including Collaborative
Filtering (CF) and Latent Factor Model (LFM). CF can be further divided into UCF
(Userbased Collaborative Filtering) and ICF (Item-based Collaborative Filtering). The core
idea of LFM is to connect users and items through latent features [
EduRank [
The most related work to our research in previous studies is the Latent Skill
Embedding (LSE) model [
Recently, reinforcement learning has been explored in personalized study planning
in ITS [
define the model based on IRT, T-IRT and MIRT model [
, ( +τ| )= , 2 ( +τ)
next moment is only relevant to his current ability value.
Where is the question discrimination,
is the question difficulty, is the
student’s ability value. The Temporal IRT (T-IRT) model [
The T-IRT model only considers interactions between students and assessments,
ignoring their interactions with learning videos. However, we believe that the students'
ability can be significantly improved after completing a learning video. Therefore, in
[
1 (−(⃗⃗ , ∙ℎ⃗⃗ ‖ℎ⃗⃗ ‖ −‖ℎ⃗⃗ ‖)) (1) (2) Where ⃗ , represents knowledge that student gains from the video , ⃗⃗ represents knowledge of the video , ℎ⃗ is the prerequisites of video , is the duration in which student watches video and is the total length of the video . In Eq (2), both student ability and learning video requirements have been expanded from one-dimensional to multidimensional. We utilize the vector projection method to determine whether the relevant abilities of the student exceed the relevant skill requirements of the video lectures.
The T-BMIRT model enables us to infer every student’s current ability , video
knowledge and video skill requirements ℎ through the student’s responses of
assessment questions. The detailed model fitting process of the T-BMIRT can be found in
[
SARLR Phase 1 describes the process of searching similar students and extracting a suitable learning path for a target student. At Step 1, the algorithm identifies the students MS with the similar skill levels to the target student through k-nearest neighbor search method over the k-dimension tree (kd-tree) structure and k-nearest neighbor search method. At Step 2-4, the algorithm selects the best student ∈ highest ability level at the moment when they complete learning specific knowledge with the units. At Step 5, the algorithm extracts the learning path of to the target student . SARLR Phase 1: Search and Extraction INPUT:
Set of students = { 1, 2, … , }, target student ∈ Matrix of abilities = [ , ], where , is the ability value of student s at time t Set of learning resources = { 1, 2, … , }
The time in this paper is the index of learning resources with the student just completed learning. OUTPUT: learning path 1: search for similar students MS, where ∈ and , 0 is similar to , 0 2: for each ∈ do 3: find = ( 4: end for 5: extract the learning path = ( 1, 2, … )of 6: return
( , − , 0)), where is the time of completing learning 3.3
(3) Because each individual student has his/her inherent learning style, even when he follows the recommended learning path generated in SARLR phase 1, the learning outcome may not be as good as expected by the recommendation algorithm. In order to deal with this problem, we set up the two conditions in Eq (3) to initiate the Adaptive Re-planning phase, which is defined in SARLR Phase 2.
We selected two datasets to perform our experiments, the public “Assistments”, including 224,076 interactions, 860 students, 1,427 assessments and 106 skills, and a blended learning data from our learning analysis platform including 14,037,146 learning behavior data from 140 schools and 9 online educational companies. 4.1
We divided each data set into two parts, one part only contains single skill assessments, and the other part contains multiple skills assessments. The IRT, T-IRT are single skill models, and the MIRT and T-BMIRT are multiple skills models. The dimensions for models are related to the numbers of knowledge components. The values in Table 1 are average results of the cross-validation. It shows that T-BMIRT outperforms the other models on each dataset, especially on the multidimensional dataset. Where ∈ is the learning resources in a recommended path, is the length of the is the knowledge components which is learning in the current chapter, function similarity() calculates the adjusted cosine similarity of the two vectors in the parentheses. The relevance accuracy
ability. The difficulty accuracy
of the recommended learning resources for the target student are matched with his is set to evaluate whether the difficulties of the
is used to evaluate whether the difficulties recommended learning resources for the target student can match his current ability levels.
We selected the blending data to do this experiments. Table 2 shows the average of
the 10-fold cross-validation results. It can be seen that the UCF and ICF have a similar
effect, but the UCF works better on the relevance accuracy, while the ICF is better at
the difficulty accuracy. The LFM performs better than the first two algorithms in terms
of both indicators. The SARLR algorithm performs best among all these algorithms.
their ability levels. We calculated “expected gain” = ( ′() − )( ) by using PCA and
K-means method to further split the students of the same group into two parts based on
their learning paths [
We developed a self-adaptive recommendation algorithm of learning resources (SARLR) to personalize students’ learning path. It contains the T-BMIRT, a temporal blended multidimensional IRT model, which performs well on the prediction task of multi-dimensional skills assessments, especially when the study process contains learning video interactions. Based on the T-BMIRT model, the SARLR algorithm adopts a reasonable recommendation strategy and establishes conditions to adaptively adjust recommendations towards the dynamic needs of the students. In addition, we extend the evaluation criteria for personalized learning recommendation in term of rationality and effectiveness. Experimental results prove that the SARLR algorithm outperforms the other recommendation algorithms based on CF and LFM.