高维数据的降维与分类——宠物消费升级的统计测度

Dimensionality Reduction and Classification of High-Dimensional Data—Statistical Measurement of Pet Consumption Upgrade

摘要: 基于近年来宠物角色从功能性饲养向情感陪伴转变的趋势，本文通过问卷调研和网络爬虫获取宠物消费的高维数据，运用交叉分析和Python词云探究宠物产品消费现状、人宠消费关系；借助因子分析进行降维，提取宠物产品消费的影响因素，并建立二元Logistic回归模型加以验证；基于K-means聚类分析将消费者划分为健康导向型、情感依赖型和性价比优先型三类。结果显示：消费者行为折射出“爱宠及物”的特征；健康养宠需求推动产品向功能化、人性化发展；消费者对宠物产品的个性化需求突显。基于此，本文为宠物产品相关经营者提出政策建议。

Abstract: Given the trend in recent years where the role of pets has shifted from functional keeping to emotional companionship, this paper obtains high-dimensional data on pet consumption through questionnaire surveys and web crawling. It employs cross-analysis and Python word clouds to explore the current state of consumption and the relationship between human and pet consumption. Factor analysis is utilized for dimensionality reduction to extract the influencing factors of pet product consumption, and a binary Logistic regression model is established for verification. Based on K-means clustering analysis, consumers are classified into three categories: health-oriented, emotionally dependent, and cost-effective priority types. The results indicate that consumer behavior reflects the characteristic of loving the pet and thus valuing related products; the demand for healthy pet keeping drives products towards functionalization and humanization; and consumers’ personalized needs for pet products are prominent. Based on these findings, this paper offers policy recommendations for operators related to pet products.