Cеминар НУЛ ММВП по теме "Frequent Itemsets and Sequences for Time Series Analysis"

В четверг 10 июля 2025 г. в 15:30 состоится онлайн семинар по теме "Frequent Itemsets and Sequences for Time Series Analysis". Докладчиком выступит Фейзал Бабер (Faisal Baber), студент магистратуры программы "Науки о данных" департамента анализа данных и ИИ ФКН. 

Аннотация: 

Financial systems keep sequential transaction data organized by time, reflecting the fact that customers use different payment methods. Detecting frequent behaviors helps spot early problems. Looking for patterns in symbolic sequences within the data allows us to see habits, notice anomalies, and create financial services based on the data. I offer a usable guide for dealing with common closed sequential patterns (FCSPs) that often appear in a real bank transaction data set. Initially, I change each raw event log into a symbolic sequence of customer actions.Using information about how much and when the transaction occurred, plus which industry the merchant belongs to. After that, I find all closed sequential pattern,non overlapping motifs in the instructional texts. According to this study, closed sequential pattern mining can be successfully used to discover patterns in transactional finance data.

 
Подключиться к конференции Zoom:

https://us06web.zoom.us/j/83977478698?pwd=jHsl32PG5FtahU3LpWW2Bhwa2Am301.1

Идентификатор конференции: 839 7747 8698
Код доступа: 365268