PhD candidate Pascan Cristian Octavian (UTCN, scientific coordinator Prof. Dr. Petrică C. Pop-Sitar) pitched his doctoral research: New Strategies and Methods for Improving Recommendation Systems.
Decision in the digital era
The act of choosing is a fundamental human trait, from antiquity to the digital era. But the digital era has produced informational overload — the absence of the time and resources needed to manually analyse the vast volume of options available to a consumer. Noise drowns out signal.
The economic relevance
35% of Amazon sales are driven by recommendation systems. Recommendation engines are at the core of success in global e-commerce, and their performance directly translates into measurable revenue. The PhD work positions itself at this intersection of consumer behaviour and algorithmic decision support.
The proposed system
The thesis proposes a recommendation system enriched with a chatbot interface — conversational assistance that helps users navigate options, refines preferences through dialogue and improves the relevance of recommendations. The chatbot brings several advantages over classical filter-based interfaces: faster intent capture, better handling of ambiguous queries and a more natural user experience.
Validation
The system is validated in an e-commerce setting, with perspectives for scaling to other application domains. The work fits into the broader COSA research line on recommendation systems — including the Trend-Enabled Recommender System with Diversity Enhancer for Crop Recommendation published by the COSA team in Agriculture (MDPI, Q1, 2025) — applied here to consumer products rather than agricultural decisions.
Presentation slides
View the full slide deck below — or open the PDF in a new tab / download.
If the PDF does not render in your browser, please use the "open" or "download" link above.
0 Comments