Abstract:

Perceived quality is the consumer’s perceptual knowledge of product attributes, which is very important to manufacturers. Considering that the product attributes perceived by consumers are different from the attribute system in enterprise production, a perceived quality attribute extraction method based on the grounded theory is constructed for online forum data. The method consists of two phases, i.e., extracting rules by inductive method process in the coding process to better adapt to the big data context, and proposing a three-stage testing process based on the text mining method in the theory saturation test to reduce the manual workload. Eight consumer-perceived car attributes are obtained in the instance analysis of a large amount of car forum data. The method ensures the completeness of attribute extraction and reduces the manual workload. This paper has theoretical significance for the research on perceived quality and attribute extraction, and improves the big data processing capability of manufacturers.

Key words:

grounded theory, online forums, attribute extraction, text mining, inductive methods

摘要：

感知质量是消费者对产品属性的感性认识，对制造企业非常重要。考虑到消费者感知到的产品属性不同于企业生产中的属性体系，面向在线论坛数据，构建了基于扎根理论的感知质量属性抽取方法。该方法由两阶段构成，首先，在编码过程中通过归纳法流程提取规则，以更好地适应大数据情境；其次，在理论饱和度检验中基于文本挖掘方法提出一个三阶段的检验流程，以降低人工工作量。在大量汽车论坛数据的实例分析中，得到了8个消费者感知到的汽车属性。该方法保证了属性抽取的完备性，并降低了人工工作量。研究对感知质量和属性抽取研究具有理论意义，也提高了制造企业的消费者大数据处理能力。

关键词:

扎根理论, 在线论坛, 属性抽取, 文本挖掘, 归纳法