Anomaly detection using one-class SVM with wavelet packet decomposition
Date
2011Anomaly detection has become a popular research topic in the field of machine learning. Support vector machine is one anomaly detection technique and it is coming one the most widely used. In this research, anomaly detection is applied to road condition monitoring, especially pothole detection, using accelerometer data. The proposed concept includes data preprocessing, feature extraction, feature selection and classification. Accelerometer data was first filtered and segmented, after which features were extracted with frequency- and time-domain functions, with genetic programming and with wavelet packet decomposition. A classification model was built using support vector machine and the calculated features. The results with actual accelerometer data demonstrates that potholes can be detected reliably. Features from wavelet packet decomposition yielded the best classification results. Poikkeavuuksien havaitsemisesta on tullut suosittu tutkimusalue koneoppimisen alalla. Tukivektorikone on yksi poikkeavuuksien havaitsemismenetelmä ja siitä on tulossa yksi alan käytetyimmistä tekniikoista. Tässä tutkielmassa poikkeavuuksien havaitsemista sovelletaan tien pinnan kuoppien tunnistamiseen kiihtyvyysanturin mittausarvoista. Kiihtyvyysanturin mittausarvoja esikäsiteltiin suodattimen ja ikkunoinnin avulla, minkä jälkeen arvoista laskettiin piirteitä aika- ja taajuustason funktioiden, geneettisen ohjelmoinnin ja aallokemuunnoksen avulla. Parhaiden piirteiden valinnan jälkeen luotiin ennustava malli tukivektorikoneella.
Luokittelutulokset osoittavat, että kuopat voidaan havaita luotettavasti kiihtyvyysanturin mittausarvoista. Parhaat tulokset saavutetiin allokemuunnoksella
lasketuilla piirteillä.
Keywords
Metadata
Show full item recordCollections
- Pro gradu -tutkielmat [29743]
License
Related items
Showing items with similar title or keywords.
-
Talent identification in soccer using a one-class support vector machine
Jauhiainen, S.; Äyrämö, S.; Forsman, H.; Kauppi, J-P. (Sciendo, 2019)Identifying potential future elite athletes is important in many sporting events. The successful identification of potential future elite athletes at an early age would help to provide high-quality coaching and training ... -
A method for anomaly detection in hyperspectral images, using deep convolutional autoencoders
Penttilä, Jeremias (2017)Menetelmä poikkeavuuksien havaitsemiseen hyperspektrikuvista käyttäen syviä konvolutiivisia autoenkoodereita. Poikkeavuuksien havaitseminen kuvista, erityisesti hyperspektraalisista kuvista, on hankalaa. Kun ongelmaan ... -
Piecewise anomaly detection using minimal learning machine for hyperspectral images
Raita-Hakola, A.-M.; Pölönen, I. (Copernicus Publications, 2021)Hyperspectral imaging, with its applications, offers promising tools for remote sensing and Earth observation. Recent development has increased the quality of the sensors. At the same time, the prices of the sensors are ... -
Feature selection in anomaly-based network intrusion detection systems
Paananen, Harri (2010) -
Unsupervised network intrusion detection systems for zero-day fast-spreading network attacks and botnets
Vahdani Amoli, Payam (University of Jyväskylä, 2015)Today, the occurrence of zero-day and complex attacks in high-speed networks is increasingly common due to the high number vulnerabilities in the cyber world. As a result, intrusions become more sophisticated and fast ...