|
KMID : 0894920210210020001
|
|
Journal of the Korean Association for Persons with Autism
2021 Volume.21 No. 2 p.1 ~ p.24
|
|
|
Analysis of Parental Perceptions on Early Symptoms of Children with Autism Spectrum Using Text Mining
|
|
Kim Myeong-Nan
Kang Hyun-Seo
Lee Yeon-Jae
Lee Mi-Ji
|
|
|
Abstract
|
|
|
|
This study analyzed parents’ perceptions of the initial symptoms of children with autism spectrum disorder using text mining. From 2017 to June 2021, 212 parental questions posted on Naver Jisik-In(intellectual people) were collected, 59 cases lacking relevance were deleted, and 153 cases were analyzed. A Python package was used for data collection, and frequency analysis was performed using Term Frequency-Inverse Document Frequency(TF-IDF) weighting technique with the first extracted words. For topic analysis, Latent Dirichlet Allocationtopic (LDA) topic modeling was performed using the ldatuning package, and the phi coefficient was used for words and topic network analysis. Subsequently, the words extracted with the highest frequency were spoken language(304 times). Furthermore, eyes (160 times), worry (158 times), behavior(125 times), hands(111 times), language (102 times), and sounds(82 times) also showed a high frequency. Five topics were analyzed, and “anxiety”, “autism”, “age” and “symptoms” were found as mediating words. word pairing and network analysis revealed that early symptoms of autism, such as “car-wheel,” “sound-shout,” and “name-call,” had a high correlation. In relational analysis by topic, related words showed a high phi coefficient under the themes of stereotypic behavior, slow speech development, developmental abnormalities(developmental regression), and parental concerns.
|
|
|
KEYWORD
|
|
|
Children with Autism spectrum, early symptoms, parental perception, text mining
|
|
|
FullTexts / Linksout information
|
|
|
|
|
|
Listed journal information
|
|
|
|