A Multimodal Retrieval-Based Framework for Popularity Prediction of Shein T-Shirts
Bibliographic record
Abstract
Popularity analysis of fashion clothing items is important, as it provides valuable insights for fashion product design and market strategy development, ultimately fostering innovation and driving economic growth. Existing approaches primarily employ computer vision techniques to extract features from clean clothing product images for predicting market response, overlooking other influential factors that contribute to a product's popularity. Moreover, since they depend on high-dimensional and abstract visual features to predict item popularity, they are prone to overfitting and offer limited explainability. To address these challenges, we propose a multimodal retrieval-based framework for predicting clothing popularity, with a case study on fast-fashion SHEIN T-shirts. To be specific, we first design a set of fashion-related questions, which are used as prompts for ChatGPT to generate structured textual descriptions for each product. These descriptions, along with the corresponding product images, serve as inputs to a multimodal retrieval model, which returns the top-k most similar popular and unpopular items for each query product. Instead of using raw high-dimensional embeddings, we calculate similarity scores between the query product and the retrieved top-k popular and unpopular items. In addition to these similarity scores, we incorporate numerical features (e.g., price, discount level) and categorical features (e.g., color etc) to build a multi-factor popularity prediction model. Experimental results demonstrate the effectiveness of the proposed framework for popularity prediction of SHEIN T-shirts and indicate the relative importance of multiple factors in shaping product popularity.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.003 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.002 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.002 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.004 | 0.002 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".