You are on page 1of 6

Comparison of Computerized and Manual

Assessment of Dyslexia Children Test using Text


Processing

Arni Muarifah Amri Adhi Dharma Wibawa Diah Puspito Wulandari


Department of Electrical Engineering Department of Computer Engineering Department of Computer Engineering
Institut Teknologi Sepuluh November Institut Teknologi Sepuluh November Institut Teknologi Sepuluh November
Surabaya, Indonesia Surabaya, Indonesia Surabaya, Indonesia
arniamri.206022@mhs.its.ac.id adhiosa@te.its.ac.id diah@te.its.ac.id

Abstract—Dyslexia is a specific learning disorder that can have an IQ above average. Miller-Medzon, (Nevid et al.)
have a negative influence, especially on children's learning states Dyslexia is 80% of cases of learning disorders and
success. Early detection and continuous therapy can help occurs in individuals who have difficulty reading even though
children achieve better success in school. Children with Dyslexia they have average intelligence.The ability to write and read is
need several evaluations and therapy during the rehabilitation an important language skill, and the occurrence of reading
programs from therapists and psychologists. The Assessment disorders causes dyslexia children to become less able to
was carried out manually and recorded using paper documents increase their potential.
until recently. Besides, the children must meet the therapist
physically. A computerized method for assessing Dyslexia Several therapy methods are very effective to overcome
children by using text processing is a new approach to better the problem of Dyslexia, including the learning methods of
record the rehabilitation progress and provide a home Orton Gillingham and Slingerland. Dyslexia children are
evaluation application so that some social distancing can also be advised to see a psychologist check the level of the weakness
done, especially in the era of the Pandemic. In this study, the of sufferers and design a suitable method of therapy. In
data was taken from 24 Dyslexia children during their three- addition, other methods are multisensory instruction therapy,
month therapy program. The assessment program is done after phonic method therapy, and visual and auditory therapy. This
the children follow the rehabilitation program from the method helps improve reading and writing skills. In addition,
therapist. During those three months, there were two times
it can also help children to catch up in school [5]. The method
scheduled for rehabilitation and two times for Assessment. The
is used at the time of therapy. Some institute courses use 1 or
Assessment uses three criteria, namely Writing Letter Series
(WLS), Arranging Sentences (AS), and Copying Sentences (CS),
2 methods of therapy for dyslexia children. In addition, some
that each Dyslexia child must do. Computer application for essential things that must be done before therapy include
testing is developed with a scoring system implemented using dyslexia children will do screening to get a diagnosis and
text processing technique. Hamming distance and Text therapy that is suitable for dyslexia children. Early therapy
similarity system were applied to score the test result and then would be better for dyslexia children.
be compared with the manual scoring method performed by the Several studies in the IT field have played many roles in
therapists. In this study, The Hamming distance was done for
developing Dyslexia, both in learning and screening-based
the criteria of the WLS, reaching an accuracy of 83% compared
applications in several areas of information technology.
to manual Assessment. For the criteria of AS, the Text
Similarity technique achieves 96% accuracy compared to the Several studies on screening and diagnosis of dyslexia
manual Assessment, and in CS reaches 88%. The results showed children have been widely conducted, including Luz Rello by
the high potential of applying text processing techniques in making screening using online gamified tests and other
assessing Dyslexia children. research from Zulhasnizam Hassan conducting research
related to screening dyslexia children using mobile games
Keywords—text mining, computerized assessment, hamming using dyslexia theory, including Rapid Automation Name
distance, text similarity, dyslexia (RAN) testing [7]. In another study conducted by Trude
Nergard, who proposed dyslexia markers for children, the test
I. INTRODUCTION was designed to identify Norwegian students who need special
According to the International Dyslexia Association, attention; computerized tests include knowledge of letters and
"Dyslexia is a specific learning disability of neurological spelling skills. The data used in this study were as many as
origin. It is characterized by difficulties in word recognition, 1000 students. The study resulted in 77% accuracy classifying
spelling ability, and letter recognition (Decoding)[1]. dyslexia children [9].
Dyslexia is considered the most common learning disability From some previous studies, the evaluation of therapy in
[2]. Dyslexia is also known as a special learning disorder that dyslexia children has not been studied more deeply. So in this
occurs in school-age. Shaywitz (2003) predicts dyslexia study, the authors tried to compare the pattern of assessment
children reach between 5% to 17% occurs in school-age results of dyslexia children before therapy and after therapy
children and, if left unchecked, will affect the development of and produce a score with dyslexia levels. This research uses
learning in school [3]. This is evidenced by the many studies assessment methods through web-based applications using
conducted by psychologists and psycholinguists who show algorithm hamming distance and text similarity.
that dyslexia children have low academic achievement even
though their intelligence level is normal and not infrequently
In computer science, information retrieval is a 3. The child is in a good mood.
fundamental problem, and text similarity and matching
algorithms are essential in implementing information 4. The child is undergoing therapy.
retrieval. One of the word matching algorithms is approximate Dyslexia data retrieval in this study used web-based
string matching, which compares text between 2 characters. applications. This application is specially designed to be easily
[8]. One algorithm for approximate string matching is understood by dyslexia children and adapted to the concept of
Hamming Distance which can evaluate similarities between dyslexia assessment that already exists. After working on the
inputs and patterns. Previous studies have implemented many assessment problem, the child's answer is saved to the
hamming distances, including spelling checkers, plagiarism database and then analyzed and matched using word
detectors, and automatic assessments. matching. Text matching uses hamming distance algorithms
In addition, one of the previous studies by Andreas about and text similarity to match the string between the child's
hamming research was "Implementation of hamming distance answer and the existing answer key. This research consists of
and brute force algorithms in detecting similarities in source several stages, including data collection, text matching using
code in the C programming language." The study resulted in hamming distance algorithm, and text similarity. Then the last
an average percentage of 20.99% better than other algorithms stage is validation and analysis. Below is a diagram of the
method in this study:

Writing Letter
Manual Scoring
Series (WLS)

Manual Arrange
Manual Scoring
Assessment Sentences (AS)

Copying Validation
Manual Scoring
Sentences (CS)
Assessment
Hamming Validation
Writing Letter
Distance
Series (AS)
Algorithm

Automatic Arrange
Text Similarity
Assessment Sentences (AS)

Copying
Text Similarity
Sentences (CS)

Fig. 1. Dyslexia Assessment Method

to detect similarities [10]. Another study, also called "tourism A. Manual Assessment Dyslexia
recommendation system using hamming distance algorithm," The dyslexia assessment aims to determine the ability of
produced 78% to provide recommendations for tourist children diagnosed with Dyslexia. Assessment is done
attractions [12]. accompanied by a psychologist or therapist about 3-4 times a
From previous research, IT-based applications for year in dyslexia children, starting from the initial Assessment
Dyslexia have not been conducted in studies related to before therapy until therapy takes place. Dyslexia children
Assessment at therapy, so in this study, the authors tried to who score far from the benchmark require the most intensive
explore and make a computerized assessment for dyslexia intervention or therapy.
children at therapy. Then the method used in this study is a The current Manual Assessment of dyslexia children is to
question and answer-based word matching method using use paper by assigning tasks to dyslexia children with the
hamming distance algorithms and text similarity. The two same criteria, namely WLS, AS, and CS. For the WLS criteria,
methods are used to compare the answers of dyslexia children the therapist mentions the letters on the problem, then the child
with the answer key. writes according to the therapist's instructions.
II. RESEARCH METHOD The therapist or psychologist accompanies only one child
at the therapy and assessment stage. This is because each child
The study consisted of primary data that consisted of the
has different difficulties and development. In addition, the
answers of 24 dyslexia children who were assessed and
therapist is also more focused on one child's Dyslexia. Some
undergoing therapy at children with special needs course. This
of the obstacles in the assessment manual include that
test consists of 3 criteria: Writing a series of letters, Arranging
therapists will need time to do manual assessments. At the
Sentences, and Copying Sentences. Assessment is applied
same time, they must also perform therapy on dyslexia
twice to see the development of evaluation results from
children. In addition, the current obstacle is that no monitoring
dyslexia children, namely before therapy and one week to 1
system helps therapists know the development of dyslexia
month after therapy.
children after Assessment at the time of therapy. It is also
In data collection, dyslexia children are confirmed to have explained in the Assessment for dyslexia reports from the
met the requirements and conditions to take the test, including: National Center for improving literacy that there is currently
no monitoring system that assesses the response of dyslexia
1. Dyslexia aged 7 years until 12 years children to the intervention and development of children at the
2. Can type on a laptop keyboard time of therapy[ 20]
B. Computerized Assessment Dyslexia measure the distance between 2 strings of the same size by
The authors tried to adopt existing manual assessments in comparing each character in the two strings. So it is easier to
this study and conduct computerized assessments. The find out whether the two strings are the same or not [12]. The
hamming distance function is suitable for using the Hamming
criteria in this study took specifications with written
distance function in the WLS criteria that calculate each
expressions related to spelling accuracy, and the accuracy of
character in the letter series.
writing sentences (APA, 2013)[3]. These criteria are then
built into the dyslexia scoring system. In this assessment Here are the stages of implementation of the hamming
system, dyslexia children will answer questions on the system distance algorithm on the WLS criteria:
by typing in front of a laptop and accompanied by a therapist.  Identify the largest string (11). For example,
The scoring system consists of three criteria: WLS, AS, and there are two strings, namely S1 and S2, then S2
CS. is the longest string. Then the formula is as
Computerized-based assessments will facilitate follows:
Assessment for dyslexia children by training in keyboard
typing, which can also strengthen knowledge of phonics and distance = length(S2) – length(S1) (1)
use muscle memory to learn word spelling[ 18]
from the above formula indicates that distance is the
C. Writing Letter Series (WLS) comparison distance between the longest string is substracted
The WLS criterion tests the ability to recognize written from the shortest string
letters from the alphabet efficiently and accurately. In this  If the string length is the same, it will be followed
criterion, dyslexia children write a series of letters based on by the hamming distance calculation as in the
what is heard from the audio on the assessment application. following matrix table e.g :
The WLS criteria in this research application consist of 3
points and have different difficulty levels. Point number 1 TABLE I. SAMPLE OF HAMMING DISTANCE
consists of 4 characters of the letter series, then in the 2nd
Answer Key A B C D
point, there are five characters of the letter series, then in the
(String 1)
3rd point, there are six characters of the letter series. All Answer by Children A P J D
characters in the letter series will be matched with the answer (String 2)
key and then scored. Value 0 1 1 0

Assessments on the WLS criteria are helpful for testing


Based on Table 1 shows two different characters, namely
orthography skills in dyslexia children. Difficulty in
in the letters B and P and then C and J. The difference is in
orthography is one of the disorders of dyslexia children that
the 2nd and 3rd order. If strings 1 and 2 are equally given a
makes it difficult for them to identify written symbols and
value of 0, while String 1 and string 2 are different, then it is
letters they hear. [15].
given a value of 1 [23].
D. Arranging Sentences (AS)
 The next stage is to determine the similarity of string
The AS criteria test a dyslexia child's ability to arrange A and string B by using the following formula:
sentences, including appropriate writing rules such as correct
spelling and grammar [25]. Dyslexia children answer Similarity: Length of the string-distance (2)
questions by arranging random words into complete sentences The formula shows that similarity is obtained from the
in this criterion. For this criterion, there are ten questions length of the string and then subtracted by the distance values
consisting of 4-12 words. The Assessment of this criterion is of S1 and S2.
to match one sentence inputted by the student with the answer
key. The order of each word will be detected by the system  After obtaining the similarity value, the next stage is
and compared to the answer key. normalized similarity. The normalized similarity is a
way of measuring the similarity of 2 strings. Here is
E. Arranging Sentences (AS) the formula used:
According to the example given, dyslexia children are 𝐷𝑖𝑠𝑡𝑎𝑛𝑐𝑒
asked to write a complete sentence at the CS criteria. Writing 𝑁𝑜𝑟𝑚𝑎𝑙𝑖𝑧𝑒𝑑 𝑆𝑖𝑚𝑖𝑙𝑎𝑟𝑖𝑡𝑦 ∶ (3)
𝐿𝑒𝑛𝑔ℎ𝑡 𝑂𝑓 𝑇ℎ𝑒 𝑆𝑡𝑟𝑖𝑛𝑔
marks and punctuation marks will also be screened on these
criteria. These criteria also include evaluating a child's The results of normalized similarity are used as scores for
orthography. dyslexia computerized assessment. The formula follows the
current manual review by calculating the correct character or
F. Hamming Distance Algorithm string and dividing by the total strings available.
Hamming Distance is one of the string matching
G. Text Similarity
algorithms used in several fields, including text mining, NLP,
and pattern recognition. Hamming distance is used to compare Text similarity is one of the techniques in text mining that
how similar the two strings are matched. [21]. When the two compares the text with available references. Text similarity
strings are equal, the distance is 0; if it is not the same, the aims to show the degree of similarity between objects [22].
distance is 1 [26]. The smaller the distance value, the more This technique is divided into four methods, as in the
similar the two strings will be [14]. following figure:
Hamming Distance algorithm is one of the popular text
matching algorithms used in string or pattern matching
applications. The advantage of hamming distance is that it can
H. Dyslexia Index Score Interpretation
Any child who has been diagnosed with Dyslexia will go
through the initial assessment stages at the time of therapy.
The results of the Assessment will determine the severity of
Dyslexia in children. The following is an Interpretation of
Dyslexia Index Scores that can help therapy or psychologists
to find out the severity of a dyslexia child [17]:

TABLE III. DYSLEXIA INDEX SCORE INTERPRETATION


Fig. 2. Four major groups of text similarity methods and algorithms
Dyslexia Index Standar Score Risk for Dyslexia
Figure 2 shows text similarity measure consists of four >98 Very low
90-97 Low
methods, including string-based, corpus-based, knowledge- 85-89 Elevated
based, and hybrid similarities. 77-84 Moderate
 String-based is a string-based match that matches 70-76 High
string sequences and character compositions <70 Very High
 Corpus-based is the matching between two concepts
based on the extraction results of the corpus. Table III shows that the higher the score of the assessment
 Knowledge-based uses representations of results, the lower the severity in dyslexia children. This index
knowledge such as facts and the meaning of words. standard is used on every criterion in the Assessment.
 Hybrid similarity: a combination of all three I. Validation
approaches of the previous type of similarity by
At the validation stage, a comparison will be carried out
adopting the advantages of string-based, corpus- between the scores of the computerized Assessment and the
based, and knowledge-based. The latest hybrid assessment manual for each criterion. The score is converted
similarity technique extracts semantic knowledge to dyslexia index interpretation to calculate the comparison
from structural wordnet representations and between computerized and manual assessments. Then the
statistical information from the internet. results of the comparison are implemented into the following
formula:
From the four types of similarity techniques described 𝑇ℎ𝑒 𝑛𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑠𝑎𝑚𝑝𝑙𝑒𝑠 𝑚𝑎𝑡𝑐ℎ𝑖𝑛𝑔
earlier, the author tried to implement text similarity-string 𝑇ℎ𝑒 𝑁𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝐴𝑙𝑙 𝑆𝑎𝑚𝑝𝑙𝑒
∗ 100% (6)
based on word matching in the AS and CS criteria. The match
is used by applying the following formula: The formula shows that the same amount of data divided
by the overall data is then multiplied by 100% and will result
𝑁𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑊𝑜𝑟𝑑𝑠 𝑀𝑎𝑡𝑐ℎ𝑒𝑑
(4) in a percentage comparison between computerized and
𝐿𝑒𝑛𝑔ℎ𝑡 𝑜𝑓 𝐾𝑒𝑦𝑤𝑜𝑟𝑑 𝐴𝑛𝑠𝑤𝑒𝑟
manual assessments.
Text similarity calculates each word and matches the III. RESULT AND DISCUSSION
order in each word in one sentence to be used as a
comparison. The advantage of the similarity-string-based text A. Comparison of Computerized Assessment and Manual
method is that it is easy to perform calculations. [24]. Assessment
Here is an example of text similarity that is run on the AS From the results of processing in computerized
and CS criteria. There are two strings, namely strings that Assessment and validated by manual Assessment. we then
contain answer keys and strings consisting of child answers: obtained the results of comparison as follows:

TABLE II. SAMPLE OF TEXT SIMILARITY TABLE IV. COMPARISON OF COMPUTERIZED ASSESSMENT
AND MANUAL ASSESSMENT
Answer “saya” “Rajin” “ke” “sekolah”
Key Criteria Matched Data Total Data of Accuracy
Answer by “saya” “racin” “sekolah” - in Dyslexia Dyslexia
Children Children Children
WLS 20 24 83%
Text similarity will operate by matching each word and AS 23 24 96%
order of words in the child's answer with the answer key.
Table 3 shows three differences between the answer and the CS 21 24 88%
answer key, namely the difference in letters in the word
"diligent" and a missing word in the phrase "to" in the child's Table IV shows the validation results of comparing
answer. So that if entered into the formula, text similarity will computerized Assessments with manual Assessments. The
produce : comparison produces a good percentage in each criterion,
2
Similarity :
4
: 0.5 (5) 83% on the WLS criteria, 96% on the US criteria, and 88%
on the CS criteria.
The above formula shows that the number of words
matching in the example is divided by the total number of
words. Then get, the similarity result is 0.5.
A. Average Result of Computerized Assessment and Manual in this computerized Assessment have higher accuracy using
Assessment hamming distance and text similarity. This result shows a
Here is the average computerized assessment score and high potential for applying computer applications for
manual Assessment on each criterion. The following average assessing the dyslexia children test during therapy. However,
score is the average of the entire children for each criterion: this study also has some shortcomings, including data
retrieval that differs between human assessment using
TABLE V. AVERAGE RESULT OF MANUAL ASSESSMENT handwriting and computerized Assessment by typing on a
AND HAMMING DISTANCE laptop.
Criteria Manual Assessment Computerized Assessment The Hamming distance was done for the criteria of the
WLS 81 76 WLS, reaching an accuracy of 83% compared to manual
AS 51 43 Assessment. For the criteria of AS, the Text Similarity
CS 69 56 technique achieves 96% accuracy compared to the manual
Assessment, and in CS reaches 88%.
And below is additional information for the average The study was limited to writing criteria in dyslexia
computerized assessment score after being converted to risk children, so it still needs further development to create more
dyslexia status: complex assessment applications for dyslexia children.
TABLE VI. RISK FOR DYSLEXIA RESULT ACKNOWLEDGMENT
Criteria Risk for Dyslexia There is no conflict of interest in writing this article. This
Manual Assessment Computerized Assessment paper was funded by the Ministry of Communications and
WLS Moderate High
Information Technology scholarship program, Republic of
AS Very High Very High
CS Very High Very High
Indonesia.
REFERENCES
From table VI, there is a difference in risk status in
[1] Roitsch, J., & Watson, S. (2019). An Overview of Dyslexia: Definition,
computerized Assessments with manual Assessments on Characteristics, Assessment, Identification, and Intervention. Science
WLS criteria. This happens because hamming distances only Journal of Education, 7(4), 81.
measure if the lengths of both strings are the same, except that https://doi.org/10.11648/j.sjedu.20190704.11
the hamming distance also does not match the sequence of the [2] el Atawy, S. M., & Ahmed, H. M. M. (2021). SPELLING CHECKER
existing letter series. Therefore, there is a difference between FOR DYSLEXIA SECOND LANGUAGE ARAB LEARNERS.
manual Assessments e.g: Journal of Theoretical and Applied Information Technology, 31(2).
www.jatit.org
String 1: abcd efg [3] Raharjo, T., & Wimbarti, S. (2020). Assessment of learning difficulties
in the category of children with Dyslexia. Jurnal Konseling Dan
String 2: abcd feg Pendidikan, 8(2), 79–85. https://doi.org/10.29210/141600
From the example above obtained results, manual [4] Journal of the International Academy for Research in Learning
Disabilities. (n.d.).
assessment produced the similarity value 4 because the
[5] Hidayati, H., Barja Sanjaya, M., Satria Pangestu, A., & Anjaini
calculation was stop after “abcd” due to the next character was Sundawa, A. (2021). Leady: A Multisensory Approach in Mobile
incorrect, while in automatic Assessment the similarity value Application for Dyslexia Children. In Communication & Information
is 5 because there was still one additional character that was Technology) Journal (Vol. 15, Issue 1).
corerct, which is “g”. We also notice that this difference [6] Rello, L., Baeza-Yates, R., Ali, A., Bigham, J. P., & Serra, M. (2020).
between the manual and computerized assessment could also Predicting risk of Dyslexia with an online gamified test. PLoS ONE,
be caused by the skill of children in using the computer 15(Dec. 12). https://doi.org/10.1371/journal.pone.0241687
keyboard, since typing, however, is different from writing. [7] Zulhasnizam Hassan, Saifuddin Mohtaram, Naim Che Pee and Abdul
Samad Shibghatullah, 2017. Dleksia Game: A Mobile Dyslexia
Mean while, other criteria in this study produce a good Screening Test Game to Screen Dyslexia Using Malay Language
comparison so that computerized Assessment can be used as Instruction. Asian Journal of Information Technology, 16: 1-6.
10.36478/ajit.2017.1.6
one of the tools for assessing dyslexia children at the time of
[8] Kim, H. J. (2021). A k-mismatch string matching for generalized edit
therapy or pre-therapy. distance using the diagonal skipping method. PLoS ONE, 16(May 5).
This study also gave results, including data on dyslexia https://doi.org/10.1371/journal.pone.0251047.
children who did therapy had different assessment results even [9] Nergård-Nilssen, T., & Friborg, O. (2021). The Dyslexia Marker Test
for Children: Development and Validation of a New Test. Assessment
at the same age. The study also resulted in dyslexia levels for Effective Intervention.
adopted from the interpretation of dyslexia index scores. https://doi.org/10.1177/15345084211063533
These results have never been reported in previous it-based [10] Budiman, A., Gunawan, D., & Hansun, S. (n.d.). Implementasi
studies and are related to dyslexia assessment applications. Algoritma Hamming Distance dan Brute Force dalam Mendeteksi
Kemiripan Source Code Bahasa Pemrograman C.Ultimatics, Vol III
IV. CONCLUSION No 2
[11] Laux, Friedrich. (2015). DBKDA 2015 the Seventh International
Computerized Assessment will further help manage Conference on Advances in Databases, Knowledge, and Data
Dyslexia in children who are in therapy. In addition, Applications, Rome, Italy, May 24 - 29, 2015. s. n.].
computerized Assessment can also help analyze the [12] Hadis, C. G., Saptono, R., & Azis, A. (n.d.). TOURISM
development of Dyslexia in children undergoing therapy or RECOMENDATION SYSTEM BY USING POSITIVE NEGATIVE
pre-therapy. The study proposes Hamming Distance and Text APRIORI AND BINARY HAMMING DISTANCE ALGORITHM.
Jurnal Ilmiah Teknologi dan Informasi. 2019.
Similarity in text mining for a more efficient assessment of
[13] Hakak, S. I., Kamsin, A., Shivakumara, P., Gilkar, G. A., Khan, W. Z.,
dyslexia children. The validation results of existing methods & Imran, M. (2019). Exact String Matching Algorithms: Survey,
Issues, and Future Research Directions. IEEE Access, 7, 69614–69637. [22] Prasetya, D. D., Wibawa, A. P., & Hirashima, T. (2018). The
https://doi.org/10.1109/ACCESS.2019.2914071 performance of text similarity algorithms. International Journal of
[14] Prakash Kalyanathaya, K., Akila, D., & Suseendren, D. (2019). A Advances in Intelligent Informatics, 4(1), 63–69.
Fuzzy Approach to Approximate String Matching for Text Retrieval in https://doi.org/10.26555/ijain.v4i1.152
NLP Network Intrusion Detection System for Cloud Computing View [23] Neha Seth (2021, Feb. 24). A simple Guide to Metrics for Calculate
project Neural Network View project A Fuzzy Approach to String Similarity. [Online]. Available:
Approximate String Matching for Text Retrieval in NLP. In Article in https://www.analyticsvidhya.com/blog/2021/02/a-simple-guide-to-
Journal of Computational Information Systems (Vol. 15). metrics-for-calculating-string-similarity/
https://www.researchgate.net/publication/333249900 [24] Wang, J., & Dong, Y. (2020). Measurement of text similarity: A
[15] Kirana, A. W. (n.d.). 2019. KEMAMPUAN MEMBACA PADA survey. In Information (Switzerland) (Vol. 11, Issue 9, pp. 1–17).
ANAK BILINGUAL DENGAN DISLEKSIA. MDPI AG. https://doi.org/10.3390/info11090421
[16] Molina, Y. (n.d.). 2016. GAMBARAN PENINGKATAN [25] State Department of Education, C. (2016). SLD/Dyslexia Assessment
PENGENALAN KATA PADA ANAK DISLEKSIA MELALUI Resource Guide. http://www.ct.gov/sde/slddyslexia
PEMBERIAN METODE SILABTIK. [26] Niranjan B Suramanian (2019). Distance / Similarity Measure in
[17] Unraveling the Mystery of Dyslexia TEDA Unraveling the Mystery of Machine Learning [online]. Available:
Dyslexia: Utilizing the Pearson Dyslexia Toolkit Unraveling the https://aiaspirant.com/distance-similarity-measures-in-machine-
Mystery of Dyslexia: Utilizing the Pearson Dyslexia Toolkit Definition learning/
of Dyslexia. (2019). [27] Taha, M. M. A., & Teuscher, C. (2020). Approximate memristive in-
[18] Meredith, Cicerchia. Touch Typing for people with Dyslexia.[Online]. memory hamming distance circuit. ACM Journal on Emerging
Available https://www.readandspell.com/touch-typing-for-dyslexias Technologies in Computing Systems, 16(2).
[19] Perera, H., Shiratuddin, M. F., Wong, K. W., & Fullarton, K. (2018). https://doi.org/10.1145/3371391
EEG Signal Analysis of Writing and Typing between Adults with
Dyslexia and Normal Controls. International Journal of Interactive
Multimedia and Artificial Intelligence, 5(1), 62.
https://doi.org/10.9781/ijimai.2018.04.005
[20] Screening for Dyslexia. 2019. National Center on Improving
Literacy.Washington, DC: U.S.2019. retrieved from
improvingliteracy.org .
[21] Maarten, Grootendorst. (2021,Feb,1) 9 Distance Measure in Data
Science. [Online]. Available: https://towardsdatascience.com/9-
distance-measures-in-data-science-918109d069fa

You might also like