Melati, Ayu Indah Shekar Melati and Karyono, Kanisius (2014) Audio detection (Audition): Android based sound detection application for hearing-impaired using AdaBoostM1 classifier with REPTree weaklearner. In: 2014 Asia-Pacific Conference on Computer Aided System Engineering, 10-12 Feb. 2014, South Kuta, Indonesia.
Full text not available from this repository.Abstract
AudiTion is an application that will help the hearing-impaired people to detect sound around them and to recognize the sound. The algorithms used in this application for Machine Learning are AdaBoostM1 functioning as a classifier and REPTree as weak learner, and it's built for Android operating system. Machine Learning is a study of computer algorithms which can improve its learning ability automatically through experience. AdaBoostM1 is one of the algorithms with Boosting method. Boosting uses all instances in each repetition, but keeping the load on any instance in the training set. REPTree is a fast decision tree learner which builds a decision/regression tree using information gain as the splitting criterion and prunes it using reduced-error pruning. Testing processes are done in four environment conditions to determine the sound prediction accuracy level. The four conditions are environments with low and high noise, far and near sound sources. AudiTion has two sound databases, the first database is indoor sounds and the second database is outdoor sounds with a total of 23 sounds. The results show that the average level of accuracy is relatively low at around 26.25% for the detection in the four conditions using both sound databases. Due to the low accuracy, we conducted trials by reducing indoor databases only for five sounds. This trial shows the accuracy of 40%. Since the accuracy results are still less than 50%, we conclude that AudiTion applications need to use another approach.
| Item Type: | Conference or Workshop Item (Paper) |
|---|---|
| Keywords: | sound detection for hearing-impaired , machine learning , AdaBoostM1 , REPTree , Android |
| Subjects: | 000 Computer Science, Information and General Works > 000 Computer Science, Knowledge and Systems > 005 Computer Programming 600 Technology (Applied Sciences) |
| Divisions: | Faculty of Engineering & Informatics > Electrical Engineering |
| Depositing User: | Administrator UMN Library |
| Date Deposited: | 08 Dec 2021 08:33 |
| Last Modified: | 08 Dec 2021 08:33 |
| URI: | https://kc.umn.ac.id/id/eprint/19393 |
Actions (login required)
 |
View Item |
