A dual-purpose system turns web clickstreams into session embeddings and LLM-distilled intent labels, delivering 1.88% Recall@1 gain and 13.38% Log Loss reduction on mobile homepage ranking plus conversion prediction improvements in financial services.
InProceedings of the 20th International Database Engineering & Applications Symposium(Montreal, QC, Canada)(IDEAS ’16)
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.IR 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
From Clicks to Intent: Cross-Platform Session Embeddings with LLM-Distilled Taxonomy for Financial Services Recommendations
A dual-purpose system turns web clickstreams into session embeddings and LLM-distilled intent labels, delivering 1.88% Recall@1 gain and 13.38% Log Loss reduction on mobile homepage ranking plus conversion prediction improvements in financial services.