利用图像衡量产品质量：CLIP Q分数及其在房地产中的应用

中文摘要 AI 辅助

CLIP Q分数是一种新颖、安全、完全可复现且计算高效的方法，用于通过对比语言-图像预训练从视觉数据中提取客观产品质量指标。我们介绍该技术并将其广泛应用于来自某在线平台的房地产数据（约50万张图像）。我们的开源指标与大语言模型（LLM）评估结果一致，且被证明是住房市场售价和租金的有力预测指标。我们还发现，更高的CLIP Q分数与更好的流动性（缩短市场停留时间）相关，尤其是对于在售房产。

英文摘要

The CLIP Q-score is a novel, safe, fully reproducible, and computationally efficient method for extracting objective product quality metrics from visual data using contrastive language-image pre-training. We introduce the technique and provide an extensive application to real estate data from an online platform ($\sim500,000$ images). Our open-source metric aligns with LLM assessments and proves to be a powerful predictor of housing market prices for both sales and rentals. We also show that a higher CLIP Q-store is associated with better liquidity (reduced time on the market), especially for properties on sale.

发表机构

• International College of Economics and Finance, HSE（高等经济大学国际经济与金融学院）
• Yandex（Yandex公司）

机构由 AI 辅助整理，请以论文原文为准。