Dengue Fever Detection using Long Short-term Memory Neural Network
© 2020 IEEE. In this research, long short-term memory is used for text classification. The LSTM model is used for the detection of dengue fever from symptoms. The inputs of the model are the text of symptoms in Thai language, as well as the sex and age of the patients. For Thai text processing, firs...
Saved in:
| Main Authors: | , , |
|---|---|
| Format: | Conference Proceeding |
| Published: |
2020
|
| Subjects: | |
| Online Access: | https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85091835865&origin=inward http://cmuir.cmu.ac.th/jspui/handle/6653943832/70427 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Institution: | Chiang Mai University |
| Summary: | © 2020 IEEE. In this research, long short-term memory is used for text classification. The LSTM model is used for the detection of dengue fever from symptoms. The inputs of the model are the text of symptoms in Thai language, as well as the sex and age of the patients. For Thai text processing, first, we will token the sentence to words, and then correct the wrong words, and convert the words to vector using Word2Vec model and set as input data for LSTM model training. In addition, we use class balanced cross-entropy loss function for solving class imbalanced data problems. The results show that the G-mean (geometric mean of the accuracy of all classes) of LSTM with class balanced cross-entropy loss of function is greater than LSTM with cross-entropy loss function. |
|---|