Peter Bakonyi


Peter Bakonyi is a Software Analyst at Aliter Technologies and a PhD student at the Faculty of Informatics and Information Technologies at STU. He possesses expertise in data engineering and machine learning.

Pythonic Outliers: Mastering LOF Detection Talk

English language

Peter Bakonyi

Join us for an insightful talk that delves into the fundamentals of leveraging the Local Outlier Factor (LOF) algorithm in Python for anomaly detection.

Anomalies, or outliers, in datasets can hold crucial information yet often remain elusive without proper tools. This presentation aims to demystify the concept of LOF and equip attendees with the knowledge to effectively identify and interpret local outliers in their own datasets.

The talk will commence with a brief overview of the theoretical underpinnings of LOF, shedding light on its ability to detect anomalies by comparing the local density of data points. We will explore the nuances of LOF, emphasizing its capacity to capture anomalies that might be overlooked by traditional methods.

We'll provide an overview of the steps involved and share practical insights into integrating LOF into data analysis workflows.

The talk will conclude by outlining considerations for parameter tuning and interpreting LOF results, providing attendees with a foundational understanding of how to apply LOF effectively in their data exploration endeavors.

By the end of the talk, attendees will have a solid understanding of the basics of LOF, its significance in anomaly detection, and the proficiency to apply this algorithm in Python. Whether you are a beginner seeking an introduction to outlier detection or an experienced practitioner looking to enhance your toolkit, this presentation promises to be a valuable exploration into the realm of local outlier factors.