=Paper=
{{Paper
|id=Vol-2753/paper48
|storemode=property
|title=A Two-step Approach in Expert Evaluation of Correctional Information Technologies for Students with Autism Spectrum Disorders
|pdfUrl=https://ceur-ws.org/Vol-2753/paper30.pdf
|volume=Vol-2753
|authors=Tetiana Shestakevych,Vasyl Andrunyk
|dblpUrl=https://dblp.org/rec/conf/iddm/ShestakevychA20
}}
==A Two-step Approach in Expert Evaluation of Correctional Information Technologies for Students with Autism Spectrum Disorders==
https://ceur-ws.org/Vol-2753/paper30.pdf
A Two-step Approach in Expert Evaluation of Correctional
Information Technologies for Students with Autism Spectrum
Disorders
Tetiana Shestakevych, Vasyl Andrunyk
Lviv Polytechnic National University, 12, Stepana Bandery Str., Lviv, Ukraine
Abstract
Consolidation and processing of the experts` knowledge in the correction of
communication and social skills of students with autism spectrum disorder seem to be
both an actual task, and a challenge. Being well-aware of special needs of each and every
student, psychologists and medics, as well as all the participants of education of ASD
students, when making a decision about which information technology to use, apply their
experience with exact ASD student. Expert evaluation technique enables support in such
decision making, and its result can be used as input information to optimization task,
solving which will allow taking into account different constraints.
Keywords 1
Autism, autism spectrum disorder, ASD, analytic hierarchy process, AHP, expert
evaluation, knapsack problem, branch and bound algorithm
1. Introduction
Personalized approach in treating the person with special needs is a key to successful his/her
socialization. An educational process seems to be the most powerful tool in improving social and
communication skills of a student. In this case, the outmost meaning an education has for students,
who have autism spectrum disorders (ASD, for general, autism). The ASD persons have difficulties
with communication, and it might be difficult to establish social connections. Education process, and
inclusion, as its best form, can be of great support in correction and improving skills of such student,
that will ensure his/her effective social life.
Supporting education of ASD students with information technologies seems to be logical (and
necessary!) development direction of inclusive education in modern life. Now It-support is widely
spreading, but unfortunately, it is unmanaged, mostly due to its complexity. Although some attempts
of its structuring were taken [1], it is still far away from being an orderly process. Psychologists,
medicians, teachers, and all the practicians, who are involved in IT support of ASD students
education, are very valuable because of their experience with the exact student, and it is crucial to find
an approach to collect their knowledge and use it to improve ASD student`s communication and
social skills. Applying expert evaluation techniques will allow involving in IT support of ASD student
education not only the abovementioned specialists but also parents, that will improve the results and
make it more personalized and student-fitted. The main objective of the research is to improve a
decision-making process for specialists who work with ASD student when then choosing the most
appropriate information technology, used to correct communication and social skills of the student
IDDM’2020: 3rd International Conference on Informatics & Data-Driven Medicine, November 19–21, 2020, Lviv, Ukraine
EMAIL: Tetiana.v.shestakevych@lpnu.ua (Tetiana Shestakevych); Vasyl.a.andrunyk@lpnu.ua (Vasyl Andrunyk)
ORCID: 0000-0002-4898-6927 (Tetiana Shestakevych); 0000-0003-0697-7384 (Vasyl Andrunyk)
©️ 2020 Copyright for this paper by its authors.
Use permitted under Creative Commons License Attribution 4.0 International (CC BY 4.0).
CEUR Workshop Proceedings (CEUR-WS.org)
(Fig. 1). To reach this goal, a combination of some approaches is proposed, i.e. considering expert
evaluation results as an optimization problem.
Figure 1. The concept of the research
2. Special needs of ASD students
The family of autism spectrum disorders is characterized by prominent biological markers, among
which are stereotyped, repetitive movements of body parts, lowered interest in social activity, lowered
gaze to faces, etc. [2]. An education process must not be a source of academic knowledge only, for
ASD students it should become a mean of improvement and correction of communication and social
skills and abilities. A balanced, thoughtful integration of information technologies into such a process
will improve it significantly, for ASD students especially. The basis of such belief ground on the fact,
that ASD student apperceive IT support and different gadgets like no other education technique [2-
16].
Education of ASD students involves specialists in different directions [3]: psychologists, teachers,
IT specialists, etc. By personalizing the process of ASD student’s education, we improve it, and
applying IT allows us to do it in a variety of different ways. For example, inclusive school can use
some information technologies for ASD students’ education. There might be some crucial questions,
like which technology is the best for the exact student? How to choose between alternative
technologies, if they differ, but slightly? How to combine the opinions of different specialists about
those technologies? How to take into account that the desirable ITs` implementation is time and
money consuming? Authors suggest an approach that allows answering answer such questions.
There are two main stages of the suggested approach of choosing the most relevant IT for ASD
student: expert evaluation and its optimization. In this research, as an expert evaluation technique, we
suggest an analytic hierarchy method (AHP). Its results, considered as a linear integer programming
task, will be solved with branch and bound algorithm.
3. Expert evaluation of information technologies for ASD
The distinguishing feature of the efficient education of students with ASD is its unconditional
personalization. That is why it is important to collect and use the knowledge of experts that work with
ASD students personally. The expert evaluation techniques are quite popular and are widely used in
different spheres, including medicine and psychology, where experts’ judgments often can be the only
available information on the issue. According to recent scientific researches, using expert evaluation
there were suggestions provide to improve the service level medical devices using Delphi method
[17]; efficiency of insurance companies was evaluated using neutrosophic data analytical hierarchy
process [18]; the implementation of strategic pathways of sustainable electricity system was analyzed,
using Aggregation of Individual Priorities method [19]; service quality evaluation system was
constructed using fuzzy analytical hierarchy process [20]; etc.
To accumulate personalized judgments of experts on the education of ASD students, the analytic
hierarchy process was chosen as experts` opinion evaluation method. The main idea of how the
method can be implemented in five consequent steps is presented at Fig. 2.
Figure 2: The concept of the AHP method
3.1 Step 1. Determine the goal
In a situation when ASD specialists should choose among many information technologies that will
help in the correction of communication and social skills, the main goal is to rank such information
technologies according to the experts` judgments.
3.2 Sep 2. Establish the decision hierarchy
All the information technologies that are to help in the correction of ASD students, should be
assessed considering their needs, and take into account personal – psychological and medical –
features of such students. The alternative information technologies should be compared according to
some criteria, which are correlated with the main goal.
Let C C i | i 1,..., n be a set of n personalized criteria, and T T j | j 1,..., m is a set of m
alternative information technologies that can be used in the correction of ASD students`
communication and social skills.
In this study, as an example, we consider n=4 personalized criteria (Table 1) and m=5 alternative
information technologies (Table 2, see [2-16]). The decision hierarchy is in fig. 3.
Table 1
Personalized criteria
Notation Criterion
C1 Communication skills improvement
C2 Social skills improvement
C3 Perception of relevant IT by the student
C4 Easement of implementation into the school system of IT support of ASD students
Table 2
Alternative IT
Notation Alternative IT
T1 Glass and augmented reality
T2 Laptop and virtual reality
T3 PECS and augmented reality
T4 Tablet and augmented reality
T5 Virtual reality
Figure 3: Decision hierarchy
3.3 Step 3. Compare the alternatives according to the criteria
There is 4 main type of experts who are close enough to the ASD student to be able to rate
alternative technologies considering personalized criteria. Such experts are school specialists, i.e.
psychologist, medic, teacher and his/her assistant, and non-school specialist, who cooperate with ASD
student – the parents. To assess the alternatives, such specialists should compare every two
alternatives considering each criterion. Fig. 4 shows how the linguistic description of the importance
can be quantified.
Figure 4: AHP rating concept
On the basis of experts` opinion, the judgment matrix A, as an m m matrix of pair-wise
comparisons is formed, A (aij ), aij 0, a ji 1 aij , aii 1 , where i 1,..., m , j 1,..., m , m is a number
of alternatives, and aij is a ration of the importance of one alternative over another (for example, fig.
5 shows the judgment matrix of one of the experts on the criterion Communication skills
mm 1
improvement). On each of n criterion, an expert should make comparisons.
2
T1 T2 T3 T4 T5
1 15 15 3 1
5 1 12 5 6
5 2 1 6 2
A
1 3 1 5 1 6 1 1 2
1 1 6 1 2 2 1
Figure 5: The pair-wise comparison matrix for Communication skills improvement criterion.
3.4 Step 4. Find relative priority for each alternative.
Experts` judgements result in n matrixes. For every matrix, each column of the matrix is
normalized, the sum of each line is divided in m, and the achieved vector w is a priority vector (fig. 6).
Accept of this method of finding the priority vector, there are some others [21-23], which differs in
precision.
T1 T2 T3 T4 T5 w
1 1 5 1 5 3 1 0.099
5 1 1 2 5 6 0.353
5 2 1 6 2 0.386
1 3 1 5 1 6 1 1 2 0.052
1 1 6 1 2 2 1 0.110
5.432, CI 0.108, CR 0.096
T1 T2 T3 T4 T5 w
1 2 4 7 4 0.417
1 2 1 2 8 7 0.322
1 4 1 2 1 5 1 0.133
1 7 1 8 1 5 1 1 2 0.041
1 4 1 7 1 2 1 0.088
5.317 , CI 0.079, CR 0.07
T1 T2 T3 T4 T5 w
1 2 1 2 1 2 1 0.143
1 2 1 1 2 1 7 2 0.093
2 2 1 1 2 2 0.209
2 7 2 1 9 0.471
1 1 2 1 2 1 9 1 0.084
5.355, CI 0.089, CR 0.079
T1 T2 T3 T4 T5 w
1 1 1 2 1 0.214
1 1 1 4 1 0.258
1 1 1 1 2 0.225
1 2 1 4 1 1 1 0.134
1 1 1 2 1 1 0.169
5.292, CI 0.073, CR 0.065
Figure 6: AHP calculations
The priority vectors form the Table 3.
Table 3
Comparison of the characteristics of five alternative information technologies
C1 C2 C3 C4 Relative priority
T1 0,099 0,417 0,143 0,214 0,218
T2 0,353 0,322 0,093 0,258 0,256
T3 0,386 0,133 0,209 0,225 0,238
T4 0,052 0,041 0,471 0,134 0,175
T5 0,110 0,088 0,084 0,169 0,113
The relative priority vector contains an overall evaluation of the alternatives, and in this research,
we calculate relative priority as average for every alternative.
3.5 Step 5. Determine the consistency of the judgements.
To determine the consistence of experts` opinions, the main eigenvalue λ is used. To find it, the
elements of each column of matrix A should be summed, and the λ can be achieved by multiplying
the found vector of sums with w. The judgements are more consistent, the closer λ is to the rank of
matrix A. To find the consistency index CI, λ should be reduced by m and divided into (m-1),
m
CI (fig. 6).
m 1
To calculate the consistency ratio CR, the random consistency index RI (Table 4, [25]) is used,
CI
according to the rank of matrix A, CR (fig. 6). The consistency ratio of less than 0.1 is considered
RI
as the fact, that the matrix meets the consistency standards [25].
Table 4
Random consistency index (M is the rank of the matrix A)
M 1 2 3 4 5 6 7 8 9
RI 0 0 0.58 0.9 1.12 1.24 1.32 1.41 1.45
According to the relative priority vector (Table 4), the most relevant information technology is T2.
No doubt, that the implementation of each information technology will have some limitation, such as
financial and time limits. Then it is possible, that some other, and not the suggested by AHP method
technology will fit better these limitations. As a suggestion, we can consider such a situation as a
knapsack problem.
4. Optimization of the set of relevant ITs as a Knapsack Problem, and branch
and bound algorithm to solve it
The Knapsack Problem (KP) is a well-known combinatorial optimization problem, introduced in
1950, and since then it overgrowth with various methods of its solving, and various spheres of its
implementation [25]. Recently, researchers formulated KP or its variations to implement it in the
Internet of Things [26]; in the optimization process of information protection tools placement [27]; in
balancing incremental revenue with financial constraints [28]; to improve the diversity of the
information exposed to social-media users, that are connected [29]; to prevent maritime cargo
disruption as during 2020 pandemic lockdowns [30]; etc. [31-35].
The mathematical programming formulation of the Knapsack Problem is:
m (1)
F ( x) z j x j Max
j 1
m (2)
aij x j bi , i 1,.., k
j 1
x j 0,1, j 1,..., m (3)
Being solved, the KP (1)-(3) gives an answer to a question, which items are packed in a knapsack
(xj=1), and which are not (xj=0). Let us formulate the KP (1)-(3) in terms of choosing the set of relevant
information technologies among available with all constraints satisfied.
Let
m=5 is the number of alternative information technologies for communication and social skills
correction (T1-T5, Table 2);
k=2 is the number of resources, the first is connected to the cost of implementation of the
technology into IT system of the school, the second one is the time for such implementation;
bi is the ith resource capacity, b1=12 money conditional units, b2=9 time conditional units;
zj is the usefulness of the jth IT (its profit), in fact, these elements are obtained from the relative
priority vector w (Table 4), where zi=1 for the least relevant alternative, and all the alternatives are evaluated
similary according to AHP results. In this case, z1=3, z2=5, z3=4, z4=2 z5=1.
aij is the amount of units of ith resource for each item j. These data are collected from technical documentation
5 7 2 8 4
on each information technology, and (a ij ) .
3 4 5 2 5
Then the KP of this research is:
F ( x) 3x1 5 x 2 4 x3 2 x 4 x5 Max (4)
5 x1 7 x 2 2 x3 8 x 4 4 x5 12 (5)
3x1 4 x 2 5 x3 2 x 4 5 x5 9
x j 0,1, j 1,...,5 (6)
To solve this task, we use the branch and bounds algorithm (Fig. 7). The resulting value will be
interpreted as a suggestion to use (the value of the variable = 1) or not to use (the value of the variable
= 0) the appropriate IT.
The KP solution is x1 0, x 2 1, x3 1, x 4 0, x5 0 . In terms of the ask of choosing the set of the
most relevant information technologies among available with time and money constraints satisfied, it is
advisable to implement the second and first of the proposed ITs, and it will take 9 money conditional units
with 9 time conditional units.
Figure 7: Application of the branch and bounds algorithm to solve (4)-(6) task
5. Conclusion
The concept of the process of choosing information technology, suitable for teaching students with autism,
consists of two steps, the expert evaluation technique, where authors suggested the analytic hierarchy
process. The second step was an optimization of the set of relevant ITs as a knapsack problem, and
branch and bound algorithm to solve it. All the methods implemented in the study were proven to be
effective, and combining them gives an additional tool in decision making for psychologists, teachers,
medicians, etc. who work with children with an autism spectrum disorder.
6. References
[1] V. Andrunyk, V. Pasichnyk, N. Antonyuk, T. Shestakevych, A complex system for teaching
students with autism: The concept of analysis. Formation of IT teaching complex, Advances in
Intelligent Systems and Computing, 2020, 1080 AISC, pp. 721-733.
[2] R.A.J. de Belen, T. Bednarz, A. Sowmya, et al., Computer vision in autism spectrum disorder
research: a systematic review of published studies from 2009 to 2019. Transl Psychiatry 10, 333
(2020). https://doi.org/10.1038/s41398-020-01015-w.
[3] T. Shestakevych, V. Pasichnyk, M. Nazaruk, , M. Medykovskiy, N. Antonyuk, Web-Products,
Actual for Inclusive School Graduates: Evaluating the Accessibility Advances in Intelligent
Systems and Computing, 2019, 871, pp. 350-363.
[4] Y. Bobalo, P. Stakhiv, N. Shakhovska, O. Hamola, Electrical Engineering Disciplines Teaching
System for Students with Special Needs, Advances in Intelligent Systems and
Computing, 2020, 938, pp. 590-599.
[5] Y. Bobalo, N. Shakhovska, P. Stakhiv, The Gamification Approach in Electrical Engineering
Disciplines Teaching for Students with Special Needs, 2019 IEEE 20th International Conference
on Computational Problems of Electrical Engineering, CPEE 2019, 2019, 8949092
[6] N. Boyko, P. Mykhailyshyn, Y. Kryvenchuk, Use a cluster approach to organize and analyze
data inside the cloud, ECONTECHMOD, 2018, Vol. 07, No. 2, pp. 15-22.
[7] V. Lytvyn, V. Vysotska, N. Shakhovska, , ... R. Peleshchak, S.Shcherbak, A smart home system
development, Advances in Intelligent Systems and Computing, 2020, 1080 AISC, pp. 804-830.
[8] M. Davydov, O. Lozynska, N. Kunanets, V. Pasichnyk, Assistive computer technologies for
people with disabilities, ECONTECHMOD, 2018, Vol. 07, No. 2, pp. 39-44.
[9] Rudovic, Ognjen & Lee, Jaeryoung & Dai, Miles & Schuller, Björn & Picard, Rosalind. (2018).
Personalized Machine Learning for Robot Perception of Affect and Engagement in Autism
Therapy. Science. 3. 10.1126/scirobotics.aao6760.
[10] Di Nuovo A, Conti D, Trubia G, Buono S, Di Nuovo S. Deep Learning Systems for Estimating
Visual Attention in Robot-Assisted Therapy of Children with Autism and Intellectual
Disability. Robotics. 2018; 7(2):25.
[11] Di Battista S, Pivetti M, Moro M, Menegatti E. Teachers’ Opinions towards Educational
Robotics for Special Needs Students: An Exploratory Italian Study. Robotics. 2020; 9(3):72.
[12] Kumazaki, Hirokazu & Muramatsu, Taro & Yoshikawa, Yuichiro & Matsumoto, Yoshio &
Ishiguro, Hiroshi & Kikuchi, Mitsuru & Sumiyoshi, Tomiki & Mimura, Masaru. (2020). Optimal
robot for intervention for individuals with autism spectrum disorders. Psychiatry and Clinical
Neurosciences. 10.1111/pcn.13132.
[13] Kversøy, Kjartan & Kellems, Ryan & Alhassan, Abdul-Razak & Bussey, Heidi & Kversøy,
Sofie. (2020). The Emerging Promise of Touchscreen Devices for Individuals with Intellectual
Disabilities. Multimodal Technologies and Interaction. 4. 10.3390/mti4040070.
[14] Tabassum, Kahkashan. (2020). Using wireless and mobile technologies to enhance teaching and
learning strategies. Indonesian Journal of Electrical Engineering and Computer Science. 17.
1555. 10.11591/ijeecs.v17.i3.pp1555-1561.
[15] Taryadi, Ichwan Kurniawan, Multimedia Augmented Reality With Picture Exchange
Communication System for Autism Spectrum Disorder, in IJCST, 2016, Vol. 7, Issue 4, p. 34.
[16] Jennifer B. Ganz, Ee ReaHong, Fara D. Goodwyn, Effectiveness of the PECS Phase III app and
choice between the app and traditional PECS among preschoolers with ASD, Research in Autism
Spectrum Disorders, 2013, Vol. 7, Issue 8, August, pp. 973–983.
[17] Zheng, Jun & Lou, Ligang & Xie, Ying & Chen, Siyao & Li, Jun & Wei, Jingming & Feng,
Jingyi. (2020). Model construction of medical endoscope service evaluation system-based on the
analysis of Delphi method. BMC Health Services Research. 20. 10.1186/s12913-020-05486-x.
[18] Z. L. Wang et al., Decision Making Methods for Evaluation of Efficiency of General Insurance
Companies in Malaysia: A Comparative Study, in IEEE Access, vol. 7, pp. 160637-160649,
2019, doi: 10.1109/ACCESS.2019.2950455.
[19] Souza Junior, Clecio & Koch, Hagen & Siegmund-Schultze, Marianna & Köppel, Johann.
(2019). An exploratory scenario analysis of strategic pathways towards a sustainable electricity
system of the drought-stricken São Francisco River Basin. Energy Systems. 1-40.
10.1007/s12667-019-00343-1.
[20] Cui, Y., Luo, F., Yang, B. et al. Construction and application of service quality evaluation system
in the preclinical research on cardiovascular implant devices. BMC Med Inform Decis
Mak 19, 37 (2019). https://doi.org/10.1186/s12911-019-0773-4
[21] Sahar Poypa, Handan urkoglu and Umit Arpacioglu, Using the analytic hierarchy process to
evaluate sustainability factors in watershed planning and management June 2020, Vol. 31, No. 1
(June 2020), pp. 78-88 Published by: Urbanistični inštitut Republike Slovenije Stable URL:
https://www.jstor.org/stable/10.2307/26928559
[22] Xianjun Qi, Mucong Zhou, Integrated energy service demand evaluation based on AHP and
entropy weight method, E3S Web Conf. 185 01046 (2020), DOI:
10.1051/e3sconf/202018501046.
[23] Ali, Ammar & Ritsema, Coen & Sayl, A GIS-Based Multicriteria Analysis in Modeling
Optimum Sites for Rainwater Harvesting. Hydrology, 2020, 7. 51. 10.3390/hydrology7030051.
[24] Tin-Chih Toly Chen, Yu-Cheng Lin, Diverse three-dimensional printing capacity planning for
manufacturers, Robotics and Computer-Integrated Manufacturing, Volume 67, 2021, 102052,
[25] Kern, Zachary & Lu, Yun & Vasko, Francis. (2020). An OR practitioner’s solution approach to
the multidimensional knapsack problem. International Journal of Industrial Engineering
Computations. 73-82. 10.5267/j.ijiec.2019.6.004.
[26] Y. Cheng, J. Zhang, L. Yang, C. Zhu and H. Zhu, "Joint Multioperator Virtual Network Sharing
and Caching in Energy Harvesting-Aided Environmental Internet of Things," in IEEE Internet of
Things Journal, vol. 7, no. 8, pp. 7689-7701, Aug. 2020, doi: 10.1109/JIOT.2020.2988321.
[27] V. A. Lakhno, M.V. Lakhno, K. T. Sauanova, Sh. N. Sagyndykova, S. A. Adilzhanova (2020).
Decision Support System on Optimization of Information Protection Tools Placement.
International Journal of Advanced Trends in Computer Science and Engineering. 9. 4457-4464.
10.30534/ijatcse/2020/39942020.
[28] D. Goldenberg J. Albert, L. Bernardi, P. Estevez, Free Lunch! Retrospective Uplift Modeling for
Dynamic Promotions Recommendation within ROI Constraints Fourteenth ACM Conference on
Recommender Systems, ACM, 2020, pp. 486-491. http://dx.doi.org/10.1145/3383313.3412215,
DOI=10.1145/3383313.3412215,
[29] Matakos, Antonis & Tu, Sijing & Gionis, Aristides. (2020). Tell me something my friends do not
know: diversity maximization in social networks. Knowledge and Information Systems. 62.
10.1007/s10115-020-01456-1.
[30] Kontovas, Christos A. and Sooprayen, Krishna, (2020), Maritime Cargo Prioritisation during a
Prolonged Pandemic Lockdown Using an Integrated TOPSIS-Knapsack Technique: A Case
Study on Small Island Developing States—The Rodrigues Island, Sustainability, 12, issue 19, pp.
1-20, https://EconPapers.repec.org/RePEc:gam:jsusta:v:12:y:2020:i:19:p:7992-:d:420392.
[31] J. Claver, A. García-Domínguez, M.A. Sebastián, Multicriteria Decision Tool for Sustainable
Reuse of Industrial Heritage into Its Urban and Social Environment. Case
Studies. Sustainability 2020, 12, 7430.
[32] A.Y. Alqahtani, A.A. Rajkhan, E-Learning Critical Success Factors during the COVID-19
Pandemic: A Comprehensive Analysis of E-Learning Managerial Perspectives. Educ.
Sci. 2020, 10, 216.
[33] Arjita Biswas, Amit Goel, Sandeep Potnis, Performance comparison of waste plastic modified
versus conventional bituminous roads in Pune city: A case study, Case Studies in Construction
Materials, Volume 13, 2020, e00411, ISSN 2214-5095,
https://doi.org/10.1016/j.cscm.2020.e00411.
[34] Singh, Rohit & Avikal, Shwetank, Selection of Best Power Supply Source for Telecom Towers
in Remote Areas. International Journal of Mathematical, Engineering and Management Sciences.
2020. 5. 913-925. 10.33889/IJMEMS.2020.5.5.070.
[35] T. Rokkas, I. Neokosmidis, Factors affecting the market adoption of cyber-security products in
energy and electrical systems: the case of SPEAR. In Proceedings of the 15th International
Conference on Availability, Reliability and Security (ARES '20). Association for Computing
Machinery, New York, NY, USA, 2020, Article 116, 1–8.
DOI:https://doi.org/10.1145/3407023.3409315.