MMoBai

How it connects

One engine, one continuous loop

The stages form a single loop, designed so that no component carries more inferential weight than it can support. Consumer voice identifies what people want and what fails them. The molecular layer ranks the pairings most compatible with those preferences. The survey closes the loop by testing whether the compatibility signal translates into real preference, and the GREEN gate confirms that it does for this panel and these concepts. The same loop can be re-run on new data and new flavour territories.

Data pipeline summary

ComponentDetails
English tweets5,021 across 6 query groups
China posts1,649 across 3 platforms (Xiaohongshu, Weibo, Douyin)
FlavorGraph nodes8,298 nodes; 8,279 screened for Variant C
Survey respondentsn = 34 (APAC urban, aged 25 to 38)

AI performance metrics

MetricValueNotes
NLP inter-model agreement (English)0.635Mean confidence, VADER vs TextBlob
NLP inter-model agreement (Chinese)0.731Mean confidence, RoBERTa-JD vs SnowNLP
Clustering k selectionk = 5Silhouette optimisation over k in 2 to 7
FlavorGraph HIGH-tier compatibility0.73 to 0.79Variant A and Variant B
FlavorGraph LOW-tier compatibility0.26Strawberry discriminative baseline
Survey validation (Spearman r)0.90 (p = 0.037)Compatibility vs mean liking, n = 34
Survey tier separationHIGH 6.9 / LOW 4.39-point hedonic scale