Contrasting Manual and Automatic Approaches for Extracting Linguistic Features in Predicting Alzheimer’s Disease through Chinese Speech
Our automatic approach for extracting linguistic features achieved comparable accuracy (88%) with those extracted through human judgment in Random Forest classifiers (92%). Importantly, MLU emerged as the predominant feature in both manually- and automatically-processed datasets. This outcome suggests a noteworthy decline in syntactic complexity during the early stages of AD in Chinese-speaking patients.
Authors
- Yi‐Hsuan Wang (ORCID: https://orcid.org/0000-0002-9170-3742)
- Yi‐Chien Liu (ORCID: https://orcid.org/0000-0003-2579-740X)
- Ya‐Ning Chang (ORCID: https://orcid.org/0000-0001-5248-0761)
- Chia‐Ying Lee (ORCID: https://orcid.org/0000-0002-1644-375X)
- Ching‐Lan Cheng (ORCID: https://orcid.org/0000-0002-7201-8491)
- Chia‐Ju Chou
Institutions
- Fu Jen Catholic University (TW)
- Tohoku University (JP)
- Institute of Linguistics, Academia Sinica (TW)
- Cardinal Tien Hospital (TW)
- National Cheng Kung University (TW)
Publication Details
- Journal
- Alzheimer s & Dementia
- Published
- 2024-12-01
- DOI
- https://doi.org/10.1002/alz.088635
- Primary Topic
- Voice and Speech Disorders
- Type
- article
- Field-Weighted Citation Impact
- 0.00