Filip Wójcik

speaker

Filip Wójcik, born in 1988, senior .NET developer, university lecturer (specialization in statistical data analysis). Fascinated with machine learning algorithms for 5 years. Worked on price modelling systems on travelling market for 2 years, currently employed in developing banking software – financial market data analysis, building recommendations and analysis of trends. Enthusiast of (widely understood) artificial intelligence, attemting to be technology-agnostic: not afraid to switch from .NET to Pyton/R/Weka tool. After work E sports fanatic, self-defense instructor and Aikido Master.

 

Presentations

Machine Learning – When Big Data Is Not Enough (2015)


Growing interest in machine learning topic in the last decade. Relation of machine learning to Big Data. Application of Machine learning to analyze Big Data Use cases (mostly from business perspective) for machine learning: Classification – customer preferences/automated expert systems construction, Pattern recognition – market basket analysis/discovering preferences/data explanation, Regression – predicting trends, modelling of processes, financial analysis, Clusterization – grouping customers/ finding similarities between customers groups, reducing the voulumes of data. Data used in machine learning: structured data – structured files/SQL data/logs, unstructured data – text, semantic web, data feeds data exploration: examining the relationships between data parts, finding correlated fatures, data transformations – changing domains, adjusting data for the concrete algorithm. Algorithms overview, large algorithms families and groups supervised learning: overview of the method, how to train such algorithms, training lifecycle and data utilization example of decision trees: step by step walkthrough for trivial example, detailed business use-cases unsupervised learning: overview of the method, how to train such algorithm, example of association rules learning, step by step walkthrough for trivial example, detailed business use-cases.