隨著資訊科技進步與網際網路蓬勃發展,具有分析價值的使用者生成內容日益增加,在龐大的數據資料中非結構化資料比重日趨上升,嘗試有別於以往的量化研究方法,針對物流服務客訴事件的議題,採取非結構化的文字探勘技術進行探討,為詳加瞭解影響客訴事件評價成因,藉由發掘未知與潛藏的顧客意見資訊,讓管理者能有效進行客訴事件評判,進行相關客訴處理流程。 本研究採用個案方式進行物流服務客訴事件之探討,藉由文字探勘技術,以情感分析方法建立一套客訴事件評價流程與分析方法,選用臺灣大學所發展的情感分析詞庫針對客訴事件進行情感標記,制定情感分析計算規則,並輔以評分系統進行情感分析運算,亦進一步取得專家評價資料,針對情感分析與專家評價進行配適度之探討。研究結果將提供快遞物流業者決策參考外,亦針對顧客觀點提供服務資源的投入配置建議,以作為學術上繼續深入研究之基礎,並將有助於顧客關係管理之發展。
With the burgeoning development of technology and the Internet, there are more and more user-generated contents that are worth analyzing. Among massive data, the ratio of the unstructured information has been increasing. Instead of taking the often-used quantitative research, this study focused on the issues on customer complaints in logistics and adopted the unstructured text mining analysis to probe deeply in order to better understand the affecting evaluation factors for customer complaints. By means of uncovering unknown and underlying truth on information for customer opinions, the managers can evaluate the customer complaint more effectively and proceed with the following handling process accordingly. This research used the case study to discover customer complaints in logistics. Via the text mining analysis, a set of processes and analyzing technique is built for customer complaints based on sentimental analysis. Sentimental analysis lexicon developed by National Taiwan University is chosen to tag emotion on customer complaints. The calculating rules for sentiment analysis is made and auxiliary rating system is used to calculate sentimental analysis. In addition, the expert appraising data is acquired to appropriately analyze the sentimental analysis and expert appraisals. The research result will offer logistics firms not only an important view on decision-making but also the suggestion for the input of resources based on customer perspectives. The result is the foundation for further academic research and it is beneficial to the development of the customer relationship management.