Service Selection Using an Ensemble Meta-Learning Classifier for Students with Disabilities

被引:1
|
作者
Namoun, Abdallah [1 ]
Humayun, Mohammad Ali [2 ]
BenRhouma, Oussama [1 ]
Hussein, Burhan Rashid [3 ]
Tufail, Ali [4 ]
Alshanqiti, Abdullah [1 ]
Nawaz, Waqas [1 ]
机构
[1] Islamic Univ Madinah, Fac Comp & Informat Syst, Madinah 42351, Saudi Arabia
[2] Informat Technol Univ, Dept Comp Sci, Lahore, Pakistan
[3] Rennes Univ, Inria, Inserm, UMR 6074,U 1228,Empenn,IRISA,ERL, Rennes, France
[4] Univ Brunei Darussalam, Sch Digital Sci, Tungku Link, BE-1410 Gadong, Brunei
关键词
service selection; disabled students; learners with disabilities; quality of service; ensemble method; deep learning; assistive technology; Internet of Things; INTERNET; THINGS; PEOPLE;
D O I
10.3390/mti7050042
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
Students with special needs should be empowered to use assistive technologies and services that suit their individual circumstances and environments to maximize their learning attainment. Fortunately, modern distributed computing paradigms, such as the Internet of Things (IoT), cloud computing, and mobile computing, provide ample opportunities to create and offer a multitude of digital assistive services and devices for people with disabilities. However, choosing the appropriate services from a pool of competing services while satisfying the unique requirements of disabled learners remains a challenging research endeavor. In this article, we propose an ensemble meta-learning model that ranks and selects the best IoT services while considering the diverse needs of disabled students within the educational context. We train and test our deep ensemble meta-learning model using two synthetically generated assistive services datasets. The first dataset incorporates 50,000 records representing the possible use of 12 learning activities, fulfilled by 60 distinct assistive services. The second dataset includes a range of 120,000 service ratings of seven quality features, including response, availability, successibility, latency, cost, quality of service, and accessibility. Our deep learning model uses an ensemble of multiple input learners fused using a meta-classification network shared by all the outputs representing individual assistive services. The model achieves significantly better results than traditional machine learning models (i.e., support vector machine and random forest) and a simple feed-forward neural network model without the ensemble technique. Furthermore, we extended our model to utilize the accessibility rating of services to suggest appropriate educational services for disabled learners. The empirical results show the acceptability of our assistive service recommender for learners with disabilities.
引用
收藏
页数:18
相关论文
共 50 条
  • [1] Using Meta-learning in the Selection of the Combination Method of a Classifier Ensemble
    da Silva, Robercy Alves
    de Paula Canuto, Anne Magaly
    Xavier Junior, Joao Carlos
    Ludermir, Teresa Bernarda
    [J]. 2018 INTERNATIONAL JOINT CONFERENCE ON NEURAL NETWORKS (IJCNN), 2018,
  • [2] On Meta-Learning for Dynamic Ensemble Selection
    Cruz, Rafael M. O.
    Sabourin, Robert
    Cavalcanti, George D. C.
    [J]. 2014 22ND INTERNATIONAL CONFERENCE ON PATTERN RECOGNITION (ICPR), 2014, : 1230 - 1235
  • [3] Automatic Recommendation Method for Classifier Ensemble Structure Using Meta-Learning
    Da Silva, Robercy Alves
    Canuto, Anne Magaly De Paula
    Barreto, Cephas Alves Da Silveira
    Xavier-Junior, Joao Carlos
    [J]. IEEE ACCESS, 2021, 9 : 106254 - 106268
  • [4] META-DES: A dynamic ensemble selection framework using meta-learning
    Cruz, Rafael M. O.
    Sabourin, Robert
    Cavalcanti, George D. C.
    Ren, Tsang Ing
    [J]. PATTERN RECOGNITION, 2015, 48 (05) : 1925 - 1935
  • [5] Evolutional Meta-Learning Framework for Automatic Classifier Selection
    Cacoveanu, Silviu
    Vidrighin, Camelia
    Potolea, Rodica
    [J]. 2009 IEEE 5TH INTERNATIONAL CONFERENCE ON INTELLIGENT COMPUTER COMMUNICATION AND PROCESSING, PROCEEDINGS, 2009, : 27 - 30
  • [6] Improving Cross-Validation Based Classifier Selection using Meta-Learning
    Krijthe, Jesse H.
    Ho, Tin Kam
    Loog, Marco
    [J]. 2012 21ST INTERNATIONAL CONFERENCE ON PATTERN RECOGNITION (ICPR 2012), 2012, : 2873 - 2876
  • [7] A Literature Survey and Empirical Study of Meta-Learning for Classifier Selection
    Khan, Irfan
    Zhang, Xianchao
    Rehman, Mobashar
    Ali, Rahman
    [J]. IEEE ACCESS, 2020, 8 : 10262 - 10281
  • [8] META-DES.Oracle: Meta-learning and feature selection for dynamic ensemble selection
    Cruz, Rafael M. O.
    Sabourin, Robert
    Cavalcanti, George D. C.
    [J]. INFORMATION FUSION, 2017, 38 : 84 - 103
  • [9] Ensemble generation and feature selection for the identification of students with learning disabilities
    Nanni, Loris
    Lumini, Alessandra
    [J]. EXPERT SYSTEMS WITH APPLICATIONS, 2009, 36 (02) : 3896 - 3900
  • [10] META-DES.H: a dynamic ensemble selection technique using meta-learning and a dynamic weighting approach
    Cruz, Rafael M. O.
    Sabourin, Robert
    Cavalcanti, George D. C.
    [J]. 2015 INTERNATIONAL JOINT CONFERENCE ON NEURAL NETWORKS (IJCNN), 2015,