Abstract
Adaptive learning offers real attention to individual students’ differences and fits different needs from students. This study proposes a bi-level recommendation system with topic models, gradient descent, and a content-based filtering algorithm. In the first level, the learning materials were analyzed by a topic model, and topic proportions to each short item in each learning material were yielded as representation features. The second level contains a measurement component and a recommendation strategy component which employ gradient descent and content-based filtering algorithm to analyze personal profile vectors and make an individualized recommendation. An empirical data consists of cumulative assessments that were used as a demonstration of the recommendation process. Results have suggested that the distribution to the estimated values in the person profile vectors were related to the ability estimation from the Rasch model, and students with similar profile vectors could be recommended with the same learning material.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
References
Amos, B., & Yarats, D. (2020). The differentiable cross-entropy method. Paper presented at the International Conference on Machine Learning.
Apaza, R. G., Cervantes, E. V., Quispe, L. C., & Luna, J. O. (2014). Online courses recommendation based on LDA. Paper presented at the SIMBig.
Beagley, J. E., & Capaldi, M. (2016). The effect of cumulative tests on the final exam. Primus, 26(9), 878–888.
Bian, L., & Xie, Y. (2010). Research on the adaptive strategy of adaptive learning system. Paper presented at the International Conference on Technologies for E-Learning and Digital Entertainment.
Blei, D. M., Ng, A. Y., & Jordan, M. I. (2003). Latent Dirichlet allocation. The Journal of Machine Learning Research, 3, 993–1022.
Cao, Y., Li, W., & Zheng, D. (2019). A hybrid recommendation approach using LDA and probabilistic matrix factorization. Cluster Computing, 22(4), 8811–8821.
Castells, P., Fernandez, M., & Vallet, D. (2006). An adaptation of the vector-space model for ontology-based information retrieval. IEEE Transactions on Knowledge Data Engineering, 19(2), 261–272.
Chen, Y., Li, X., Liu, J., & Ying, Z. (2018). Recommendation system for adaptive learning. Applied Psychological Measurement, 42(1), 24–41.
Cheng, Y., & Bu, X. (2020). Research on key technologies of personalized education resource recommendation system based on big data environment. Paper presented at the Journal of Physics: Conference Series.
Chou, Y.-T., & Wang, W.-C. (2010). Checking dimensionality in item response models with principal component analysis on standardized residuals. Educational and Psychological Measurement, 70(5), 717–731.
Cleghorn, G. D. (1986). Policies of US medical schools on the use of the NBME Part I and Part II examinations. Journal of Medical Education, 61(12), 954–957.
De Boer, P.-T., Kroese, D. P., Mannor, S., & Rubinstein, R. (2005). A tutorial on the cross-entropy method. Annals of Operations Research, 134(1), 19–67.
den Boer, A. W., Verkoeijen, P. P., & Heijltjes, A. E. (2021). Comparing formative and summative cumulative assessment: Two field experiments in an applied university engineering course. Psychology Learning & Teaching, 20(1), 128–143.
Dhawan, S. (2020). Online learning: A panacea in the time of COVID-19 crisis. Journal of Educational Technology Systems, 49(1), 5–22.
Engelhard, G. Jr. (2013). Invariant measurement: Using Rasch models in the social, behavioral, and health sciences. Routledge.
Georgia Center for Assessment. (2018). English language and Arts cumulative Assessment.
Ghauth, K. I., & Abdullah, N. A. (2010). Learning materials recommendation using good learners’ ratings and content-based filtering. Educational Technology Research and Development, 58(6), 711–727.
Imhof, C., Bergamin, P., & McGarrity, S. (2020). Implementation of adaptive learning systems: Current state and potential. In Online teaching and learning in higher education (pp. 93–115). Springer.
Koedinger, K. R., Brunskill, E., Baker, R. S., McLaughlin, E. A., & Stamper, J. (2013). New potentials for data-driven intelligent tutoring system development and optimization. AI Magazine, 34(3), 27–41.
Kuang, W., Luo, N., & Sun, Z. (2011). Resource recommendation based on topic model for educational system. Paper presented at the 2011 6th IEEE Joint International Information Technology and Artificial Intelligence Conference.
Liang, Q., & Hainan, N. C. (2019). Adaptive learning model and implementation based on big data. Paper presented at the 2019 2nd International Conference on Artificial Intelligence and Big Data (ICAIBD).
Lin, Q., He, S., & Deng, Y. (2021). Method of personalized educational resource recommendation based on LDA and learner’s behavior. The International Journal of Electrical Engineering & Education, 0020720920983511.
Mavroudi, A., Giannakos, M., & Krogstie, J. (2018). Supporting adaptive learning pathways through the use of learning analytics: Developments, challenges and future opportunities. Interactive Learning Environments, 26(2), 206–220.
Mei, Q., Shen, X., & Zhai, C. (2007). Automatic labeling of multinomial topic models. Paper presented at the Proceedings of the 13th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining.
Pfennig, A. (2020). Improving learning outcome for GSL (German as a Second Language) students in a blended learning cumulative assessment material science course. Paper presented at the Int. Conf. on Education and E-Learning ICEEL 2020.
Romero, C., Ventura, S., Delgado, J. A., & De Bra, P. (2007). Personalized links recommendation based on data mining in adaptive educational hypermedia systems. Paper presented at the European conference on technology enhanced learning.
Ryan, G. J., & Nykamp, D. (2000). Use of cumulative examinations at US schools of pharmacy. American Journal of Pharmaceutical Education, 64(4), 409–412.
Tang, X., Chen, Y., Li, X., Liu, J., & Ying, Z. (2019). A reinforcement learning approach to personalized learning recommendation systems. British Journal of Mathematical and Statistical Psychology, 72(1), 108–135.
Tong, Z., & Zhang, H. (2016). A text mining research based on LDA topic modelling. Paper presented at the International Conference on Computer Science, Engineering and Information Technology.
United States Department of Education. (2017). Reimagining the role of technology in education: 2017 National Education Technology Plan update. Washington, DC Retrieved from https://tech.ed.gov/files/2017/01/NETP17.pdf
USMLE. (2014). Federation of State Medical Boards of the United States and the National Board of Medical Examiners. USMLE Bulletin of Information.
Wheeler, J. M., Cohen, A. S., Xiong, J., Lee, J., & Choi, H.-J. (2021). Sample size for latent Dirichlet allocation of constructed-response items. In Quantitative Psychology (pp. 263–273). Springer.
Xiong, J., Choi, H.-J., Kim, S., Kwak, M., & Cohen, A. S. (2019). Topic Modeling of Constructed-Response Answers on Social Study Assessments. In The annual meeting of the psychometric Society (pp. 263–274). Springer, Cham.
Yan, X., Guo, J., Lan, Y., & Cheng, X. (2013). A biterm topic model for short texts. Paper presented at the Proceedings of the 22nd International Conference on World Wide Web.
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Appendix
Appendix
1.1 Data
Cumulative Assessments are aligned and assess a representation of the Georgia Standards of Excellence (GSE). These cumulative forms can help teachers to gather strong evidence on student learning toward the overall end-of-year expectations at each grade level. The sample assessment items were provided on this web site: https://www.lennections.com/assesslets-science
1.2 R Code
############################################################ ##########################read data######################### ############################################################ data<-read.csv(file='data.csv', header=T, sep=",", fill=T, stringsAsFactors = F) #processing data2 <- udpipe(data, "english") biterms <- as.data.table(data2)[, cooccurrence(x = lemma, relevant = upos %in% c("NOUN", "ADJ", "VERB") & nchar(lemma) > 2 & !lemma %in% stopwords("en"), skipgram = 3), by = list(doc_id)] data3 <- data2[, c("doc_id", "lemma")] ############################################################ ####################decide optimal numbers################## ############################################################ cd_k<-seq(2,10) #JSD model=NULL for (i in cd_k) { model[[i]] <- BTM(data3, biterms = biterms, k = i, alpha = 1, beta = 1, window = 3, iter = 5000, background = F, trace = F,detailed = F) } # Compute Jensen-Shannon Divergence for each value in model scores <- predict(model[[1]], newdata = data3) colnames(scores)<-c("topic1","topic2","topic3","topic4") JSD <- function(p, q) { m <- 0.5 ∗ (p + q) divergence <- 0.5 ∗ (sum(p ∗ log(p / m)) + sum(q ∗ log(q / m))) return(divergence) } n <- dim(scores)[1] X <- matrix(rep(0, n∗n), nrow=n, ncol=n) indexes <- t(combn(1:nrow(scores), m=2)) for (r in 1:nrow(indexes)) { i <- indexes[r, ][1] j <- indexes[r, ][2] p <- scores[i, ] q <- scores[j, ] X[i, j] <- JSD(p,q) } ############################################################ ####################Estimation and predict################## ############################################################ #read students' response data student = read.csv(file='4_Cumulative_Assesslet.csv', header=T, sep=",", fill=T,stringsAsFactors = F) #M4 is feature matrix of 4th assessment M4 = read.csv(file='M4.csv', header=T, sep=",", fill=T,stringsAsFactors = F) X = as.matrix(M4) y= as.matrix(student) N= dim(y)[1]∗dim(y)[2] theta.init = matrix(rnorm(n=dim(X)[2]∗dim(y)[1], mean=0,sd = 1), nrow=dim(y)[1],ncol=dim(X)[2], byrow=T) e = y - theta.init%∗%t(X) grad.init = -(2/N)∗(e)%∗%X theta = theta.init - eta∗(1/N)∗grad.init l2loss = c() for(i in 1:iters){ myMatrix = y - theta%∗%t(X) # empty matrix for the results squaredMatrix = matrix(nrow=dim(myMatrix)[1], ncol=dim(myMatrix)[2]) for(i in 1:nrow(myMatrix)) { for(j in 1:ncol(myMatrix)) { squaredMatrix[i,j] = myMatrix[i,j]ˆ2 } } l2loss = c(l2loss,sqrt(sum(squaredMatrix))) e = y - theta%∗%t(X) grad = -(2/N)∗e%∗%X theta = theta - eta∗(2/N)∗grad # empty matrix for the results squaredMatrix2 = matrix(nrow=dim(grad)[1], ncol=dim(grad)[2]) for(i in 1:nrow(grad)) { for(j in 1:ncol(grad)) { squaredMatrix2[i,j] = grad[i,j]ˆ2 } } if(sqrt(sum(squaredMatrix2)) <= epsilon){ break } } values<-list("coef" = theta, "l2loss" = l2loss) h=sigmoid(X%∗%t(theta.init)) sum(diag(-y%∗%log(h)-(1-y)%∗%log(1-h)))/m #sigmoid function, inverse of logit sigmoid <- function(z){1/(1+exp(-z))} #initialize theta theta <- matrix(rnorm(n=dim(X)[2]∗dim(y)[1], mean=0,sd = 1), nrow=dim(y)[1],ncol=dim(X)[2], byrow=T) #comput GD compCost<-function(para){ m <- dim(y)[1]∗dim(y)[2] j=0 for (i in seq(1,492∗4,by=4)) { k=match(i,seq(1,492∗4,by=4)) l1_1=sigmoid(colSums(para[i:(i+3)]∗t(X))) l1 <- log(l1_1) l2 <- log(1-l1_1) j=j+sum(y[k,]∗l1+(1-y[k,])∗l2) } J=-j/m }
Rights and permissions
Copyright information
© 2022 The Author(s), under exclusive license to Springer Nature Switzerland AG
About this paper
Cite this paper
Xiong, J., Wheeler, J.M., Choi, HJ., Cohen, A.S. (2022). A Bi-level Individualized Adaptive Learning Recommendation System Based on Topic Modeling. In: Wiberg, M., Molenaar, D., González, J., Kim, JS., Hwang, H. (eds) Quantitative Psychology. IMPS 2021. Springer Proceedings in Mathematics & Statistics, vol 393. Springer, Cham. https://doi.org/10.1007/978-3-031-04572-1_10
Download citation
DOI: https://doi.org/10.1007/978-3-031-04572-1_10
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-031-04571-4
Online ISBN: 978-3-031-04572-1
eBook Packages: Mathematics and StatisticsMathematics and Statistics (R0)