Dementia Detection Using LSTM and GRU


  • Neha Shivhare
  • Shanti Rathod
  • M. R. Khan



Dementia Detection, Long Short-Term Memory, GRU, Speech Analysis, features extraction.


Neuro-degenerative infections, like dementia, can affect discourse, language, and the ability of correspondence.
A new report to work on the precision of dementia identification examined the utilization of conversation analysis
(CA) of meetings between patients and nervous system specialists to recognize reformist neuro-degenerative
(ND) memory issues patients and those with (non-reformist) FMD (Functional Memory Disorder). In any case,
manual CA is expensive for routine clinical use and hard proportional. In this work, we present an early dementia
discovery framework utilizing discourse acknowledgment and examination dependent on NLP method and
acoustic component handling strategy apply on various element extraction and learning using LSTM (Long
Short-Term Memory) and GRU which strikingly catches the transient provisions and long haul conditions from
authentic information to demonstrate the abilities of grouping models over a feed-forward neural organization in
estimating discourse investigation related issues. Dementia dataset is taken where the audio file is considered for
speech recognition analysis on basis of that data is generated and it is predefined given in dementia data databank.
That audio file is converted to text based on speech analysis. Using LSTM and GRU gives efficient results.




How to Cite

Neha Shivhare, Shanti Rathod, & M. R. Khan. (2022). Dementia Detection Using LSTM and GRU. JOURNAL OF ADVANCED APPLIED SCIENTIFIC RESEARCH, 4(1).