Content uploaded by Ngo Le Huy Hien
Author content
All content in this area was uploaded by Ngo Le Huy Hien on Apr 02, 2020
Content may be subject to copyright.
A preview of the PDF is not available
Artificial Neural Network and Fuzzy Logic Approach to diagnose Autism Spectrum Disorder
Abstract and Figures
Autism Spectrum Disorder (ASD) is becoming a big issue in numerous countries around the world which can even negatively affect human natural evolution. Even though autism can be diagnosed early-before 2 years old, most children were not diagnosed with ASD until the age of 4 because of its complex symptoms and ambiguous manifestation in infant's disorders. Applying science and technology into early autism diagnosis is of vital importance, especially when data mining branches and decision-making support systems are developing and achieving many accomplishments in various fields, medicine included. Contributing to those developments, the combination between the Artificial Neural Network (ANN) and Fuzzy logic has triggered a huge revolution in data mining and is able to solve a variety of problems. This paper is the elaboration on the method of employing this combination to facilitate the early diagnosis of ASD. The result of the paper shows that the aforementioned approach has the potential to be the fundamental basis of the supporting decision-making system in ASD researching and diagnosing.
Figures - uploaded by Ngo Le Huy Hien
Author content
All figure content in this area was uploaded by Ngo Le Huy Hien
Content may be subject to copyright.
... The main goal of the studies that were involved with artificial intelligence and its applications in the diagnosis of autism was to improve the accuracy of the existed classification scales but, also, to create new scales based on neural networks [24]. Inspired by the human brain the researchers tried to develop algorithms that represent certain levels of intelligence. ...
... Nguyen and Ngo at their research suggest a combination of neural networks and fuzzy logic for the early diagnosis of autism [24]. Since the artificial neural networks demand a big amount of samples and due to the difficulty to find real data, they decided to use a temporary database that was created for CARS (Childhood Autism Rating Scale). ...
... Since the artificial neural networks demand a big amount of samples and due to the difficulty to find real data, they decided to use a temporary database that was created for CARS (Childhood Autism Rating Scale). For the relationship of the child with people, four levels were created that show the level of relevance with autism [24]. Level number 1 states that there are not elements of difficulty with the relationship with others, level number 2 that the relationships of the child are slightly dysfunctional, level number 3 that the relationships are relatively dysfunctional while level number 4 states seriously dysfunctional relationships. ...
The current paper review gives a brief and representative description of the role that artificial intelligence plays nowadays at the assessment of autism. Therefore, many researchers note that artificial intelligence plays a notable role in the early diagnosis of autism since it helps the clinicians shorten the diagnose process and have more accurate results. Thus, the research team of this paper presents some applications of artificial intelligence that are used already or are in a preliminary phase aiming to highlight the use of smart technology in the diagnosing process of autism. Lastly, it is worth noting that an early and accurate diagnose is the key point for an individualized and successful intervention which aids the academic as well as the personal de-velopment of the child.
... Our AI-based data analytics to identify ASD on the basis of the children's narrative and vocabulary skills may be viewed as part of a broader AI approach to improving objectivity in the early diagnosis of ASD, as well as to enhancing access to clinical services and educational opportunities for these individuals. AI-based designs have so far mainly capitalized on cognitive and behavioral phenotypes in children with ASD, including stereotypical behaviors [74], diagnostic measures such as the Autism Diagnostic Observation Schedule (ADOS) [75], the Autism Diagnostic Interview-Revised (ADI-R) [54], or the Childhood Autism Rating Scale (CARS) [76][77][78]. Interestingly, ML classifiers have been recently developed to predict which types of teacher communication strategies are more likely to generate positive responses in seven students with ASD in the classroom [79]. ...
Despite the consensus that early identification leads to be er outcomes for individuals with autism spectrum disorder (ASD), recent research reveals that the average age of diagnosis in the Greek population is approximately six years. However, this age of diagnosis is delayed by an additional two years for families from lower-income or minority backgrounds. These disparities result in adverse impacts on intervention outcomes, which are further burdened by the often time-consuming and labor-intensive language assessments for children with ASD. There is a crucial need for tools that increase access to early assessment and diagnosis that will be rigorous and objective. The current study leverages the capabilities of artificial intelligence to develop a reliable and practical model for distinguishing children with ASD from typically-developing peers based on their narrative and vocabulary skills. We applied natural language processing-based extraction techniques to automatically acquire language features (narrative and vocabulary skills) from storytelling in 68 children with ASD and 52 typically-developing children, and then trained machine learning models on the children's combined narrative and expressive vocabulary data to generate behavioral targets that effectively differentiate ASD from typically-developing children. According to the findings, the model could distinguish ASD from typically-developing children, achieving an accuracy of 96%. Specifically, out of the models used, hist gradient boosting and XGBoost showed slightly superior performance compared to the decision trees and gradient boosting models, particularly regarding accuracy and F1 score. These results bode well for the deployment of machine learning technology for children with ASD, especially those with limited access to early identification services.
... and discussion of data used in training AIOur research has identified several types of data that are commonly used in the training of AI. Firstly, questionnaires were used as a type of text data, with examples including AQ-10 (Autism Spectrum Quotient) (Abdullah et al., 2019; Akter et al., 2019; Elavarasi et al., 2020; Erkan & Thanh, 2019; Guimarães et al., 2019; Halibas et al., 2018; Omar et al., 2019; Peral et al., 2020; Raj & Masood, 2020; Thabtah & Peebles, 2020; Usta et al., 2019; Vaishali & Sasikala, 2018), ADI-R (The Autism Diagnostic Interview -Revised)(Bone et al., 2016;Kohli et al., 2022;Wall, Dally, et al., 2012), ADOS (Autism Diagnostic Observation Schedule)(Kosmicki et al., 2015;Levy et al., 2017;Narzisi et al., 2015;Wall, Kosmicki, et al., 2012), MARA (The Mobile Autism Risk Assessment)(Duda et al., 2016), CARS (Childhood Autism Rating Scale)(Nguyen & Le Huy Hien, 2018), Skills assessment(Stevens et al., 2019), MSEL (Mullen Scales of Early Learning)(Bussu et al., 2018) and SRS (The Social Responsiveness Scale) ...
We declare that we have no involvement in the commercialization of any products or services aimed for use within the scope of this project, and we have not received funding from any public or private organizations. Our only intention is to provide our proposed "AAA" (accessible affordable assessment) solution to those in need without any hidden motives. The design of a project with such a large-scale and complex technical and organizational requirement necessitates cooperation with various partners and institutions. The sole purpose of this document is to present our findings and suggestions. Disclaimer 2: Data protection issue. Throughout the project we will follow relevant regulations and standards, including: GDPR, HIPAA, HL7 and any relevant local regulations and standards. A detailed Data Management Plan (DMP) will be written for the project. It will explain how all relevant regulations will be followed on a practical level. DMP will include information regarding data collection methods, data storage solutions and data access control. It would also state how ethical issues such as confidentiality, anonymization and allowed data use would be addressed in practice at all stages of the project. 3 Abstract Access to autism assessment or to information supporting self-identification as autistic is limited worldwide due to an insufficient number of trained clinicians, high cost of formal diagnosis, underdeveloped health services and difficulties in obtaining reliable information about autism for those who choose to seek such information. Access to these assessments as an adult or female can be especially difficult due to outdated understanding of autism. Benefits of assessment include a better understanding of one's strengths and needs as well as the opportunity to obtain adequate support. This in turn may improve one's wellbeing or support one's employment. New technologies such as Artificial Intelligence (AI) and Internet of Things (IoT) can potentially offer solutions to overcome current barriers in access to formal diagnosis or support in self-identification. If successful, AI can help in automation of the autism assessment process. With the technology development and adoption, cost of devices collecting assessment data should become negligible. The use of technology would make access to the relevant services possible, even in the most remote locations. In this paper, we review recent evidence on the use of AI in autism research, discuss the strengths and limitations of different AI techniques, and explore future directions for research in this field. We present a summary of data types used to train AI and discuss issues related to their quality. This is followed by an overview of devices to collect data and a suggestion of KIT 1 to be used in future research. Based on the evidence and experts' opinion, we formulated a framework for building accessible and affordable autism assessment tools. Following the discussion of gaps in current evidence, we propose a "1000 people" project, in which we aim to create a comprehensive data repository for AI training with the goal of developing the tool. Our report closes with a list of questions to expert panels, which will inform further development of the project.
Autism is one of those psychiatric disorders that affect an individual's social, personal, and professional sphere(s). Autism, especially in children, is one of the most common behavioral disorders, wherein lack of good communication understandability exists throughout adulthood. The prescribed treatment of an autistic child depends completely upon an exhaustive, accurate examination of the child, which is very difficult for clinicians and consultants. Since autism is a complicated and very difficult psychiatric disorder, clinicians have joined hands with computational biologists to solve the foundations underlying the detection and diagnosis of autism. There is enormous literature evidence that highlights the pros of using machine learning techniques to develop an efficient, accurate, and robust autism detection and diagnosis system. Based on soft computing approaches, many researchers have proposed fuzzy logic-based solutions for modeling and predicting autism spectrum disorder. This chapter highlights the current scenario of autism detection and diagnosis, and how a soft computing-based intelligent agent—fuzzy logic systems—is being used to predict autism and its grades in children.
Breast cancer is the second cause of dead among women. Early detection followed by appropriate cancer treatment can reduce the deadly risk. Medical professionals can make mistakes while identifying a disease. The help of technology such as data mining and machine learning can substantially improve the diagnosis accuracy. Artificial Neural Networks (ANN) has been widely used in intelligent breast cancer diagnosis. However, the standard Gradient-Based Back Propagation Artificial Neural Networks (BP ANN) has some limitations. There are parameters to be set in the beginning, long time for training process, and possibility to be trapped in local minima. In this research, we implemented ANN with extreme learning techniques for diagnosing breast cancer based on Breast Cancer Wisconsin Dataset. Results showed that Extreme Learning Machine Neural Networks (ELM ANN) has better generalization classifier model than BP ANN. The development of this technique is promising as intelligent component in medical decision support systems.
Despite the significance of Artificial Neural Network (ANN) algorithm to
market segmentation, there is a need of a comprehensive literature review and a
classification system for it towards identification of future trend of market
segmentation research. The present work is the first identifiable academic
literature review of the application of neural network based techniques to
segmentation. Our study has provided an academic database of literature between
the periods of 2000-2010 and proposed a classification scheme for the articles.
One thousands (1000) articles have been identified, and around 100 relevant
selected articles have been subsequently reviewed and classified based on the
major focus of each paper. Findings of this study indicated that the research
area of ANN based applications are receiving most research attention and self
organizing map based applications are second in position to be used in
segmentation. The commonly used models for market segmentation are data mining,
intelligent system etc. Our analysis furnishes a roadmap to guide future
research and aid knowledge accretion and establishment pertaining to the
application of ANN based techniques in market segmentation. Thus the present
work will significantly contribute to both the industry and academic research
in business and marketing as a sustainable valuable knowledge source of market
segmentation with the future trend of ANN application in segmentation.
Artificial neural networks are more and more used for classification. They are capable to process incomplete and imprecise data and to detect non-linear relations in the data. We describe the application of neural networks for classification in the area of geology. An unsupervised learning neural network system was applied to the classification of biogene sedimentation. The classification by neural networks performed better than statistical cluster analysis. In addition, a rule generation algorithm extracted rules out of the neural network, which could be used by the geological expert. Keywords Neural network, classification, biogene sedimentation, rule generation 1.
Artificial neural networks are the machine learning technique best known in the high-energy physics community. Introduced in the field in 1988, followed by a decade of tests and applications received with reticence by the community, they became a common tool in high-energy physics data analysis. Important physics results have been extracted using this method in the last decade. This lecture makes an introduction of the topic discussing various types of artificial neural networks, some of them commonly used in high-energy physics, other not explored yet. Examples of applications in high-energy physics are also briefly discuss with the intention of illustrating types of problems which can be addressed by this technique rather than providing a review of such applications.
Follow-up Interview for the Modified Checklist for Autism in Toddlers (M-CHAT FUI). Self-published. Acknowledgement: We thank Joaquin Fuentes, M.D. for his work in developing the flow chart format used in this interview.
Autism and Developmental Disabilities Monitoring Network
Autism and Developmental Disabilities Monitoring
Network, Community report on Autism 2016, 2016.
- Wikipedia
Wikipedia, Childhood Autism Ratin Scale definition,
https://en.wikipedia.org/wiki/Childhood_Autism_Rat
ing_Scale, Dec. 2017.