Date: 09/01/2023
Hour:11.00
Student Name & ID: Merve KURU & 215105101
Supervisor: Assoc. Prof. Ömer KARAL
Topic: Incremental Local Outlier Probabilities
Link or Room: C401
Abstract:
“The rapid development of computer technology and the ease of access to it cause the emergence of different types of data. However, in these produced data, besides meaningful and useful information, undesirable data points (outliers) can also be observed due to dynamic environmental conditions. Such situations can lead to erroneous results or wrong decisions. Therefore, the discovery of outliers in many real-time applications is still one of the most important research topics. In this study, an exact instance incremental LOOP (Local outlier probability) algorithm was developed to detect outliers from real-time data streams. The developed iLOOP algorithm is compared with existing incremental outlier algorithms in detail and its advantages are reported.”
