| DC 欄位 |
值 |
語言 |
| DC.contributor | 資訊管理學系 | zh_TW |
| DC.creator | 蔡名宣 | zh_TW |
| DC.creator | Ming-Hsuan Tsai | en_US |
| dc.date.accessioned | 2018-7-10T07:39:07Z | |
| dc.date.available | 2018-7-10T07:39:07Z | |
| dc.date.issued | 2018 | |
| dc.identifier.uri | http://ir.lib.ncu.edu.tw:88/thesis/view_etd.asp?URN=105423044 | |
| dc.contributor.department | 資訊管理學系 | zh_TW |
| DC.description | 國立中央大學 | zh_TW |
| DC.description | National Central University | en_US |
| dc.description.abstract | 網路口碑是許多人搜尋醫療服務的重要依據，但網路資料量龐大，要找出符合需求的資訊相當耗時，推薦系統 (Recommender) 是解決上述問題的重要方法之一。在推薦系統設計中，冷啟動 (Cold-Start) （即一資料集中，若某項目從未被使用者評分時，要推薦此項目就變得困難）是個急待解決的問題，因為它會顯著降低推薦準確度。為此，本研究設計一套推薦機制，結合網路口碑與Anderson醫療利用理論 (Andersen Healthcare Utilization Model) 建立一套結合協同過濾式 (Collaborative filtering, CF) 及基於內容方法 (Content-based filtering, CBF) 的混合式牙醫推薦系統 (名為Dental Care Recommender, DCR)，目的在解決冷啟動並協助使用者更快做出適當決策。
為了解DCR的實用性與精確度，本研究招募50位受測者使用DCR和Google搜尋引擎，並比較二者差異。實驗結果顯示DCR的推薦精確度顯著高於Google。而使用者對DCR的搜尋時間、系統品質、系統效能、系統觀感與使用意願的滿意度也較Google優異 (p<0.05)。本研究結果可提供醫療服務業者、管理者做為重要參考。 | zh_TW |
| dc.description.abstract | Many people now search for medical services based on electronic word-of-mouth (eWOM). However, with the vast amount of information existing on the internet, finding the information that meets one’s needs is extremely time-consuming, and recommender systems constitute one of the important solutions to this issue. In the design of recommender systems, cold start (in which an item in a dataset is difficult to recommend when said item has not been rated by users) is a problem in urgent need of a solution because it significantly reduces the accuracy of recommendations. In view of this, we used eWOM and the Andersen Healthcare Utilization Model to develop a hybrid recommendation mechanism combining collaborative filtering (CF) and content-based filtering (CBF). The objective of our system, called the Dental Care Recommender (DCR), is to solve the cold-start problem and help users in making good decisions quickly.
To understand the usefulness and accuracy of the DCR, we recruited 50 participants to use the DCR and the Google search engine and compare the two. The results indicated that the recommendations made by the DCR were significantly more accurate than those made by Google. The participants also expressed significantly greater satisfaction with the search time, system quality, system performance, and system perception in the DCR than Google and significantly greater willingness to use the DCR (p<0.05). The results of this study can provide crucial reference to medical service providers and managers. | en_US |
| DC.subject | 推薦系統 | zh_TW |
| DC.subject | 混合式推薦 | zh_TW |
| DC.subject | 網路口碑 | zh_TW |
| DC.subject | 冷啟動問題 | zh_TW |
| DC.subject | 語意分析 | zh_TW |
| DC.subject | 安德森醫療利用理論 | zh_TW |
| DC.subject | recommender system | en_US |
| DC.subject | hybrid recommendation | en_US |
| DC.subject | eWOM | en_US |
| DC.subject | cold-start problem | en_US |
| DC.subject | semantic analysis | en_US |
| DC.subject | Andersen Healthcare Utilization Model | en_US |
| DC.title | 基於網路口碑與醫療利用理論之混合式推薦系統 | zh_TW |
| dc.language.iso | zh-TW | zh-TW |
| DC.title | Hybrid Recommender System Based on eWOM and Healthcare Utilization Theory | en_US |
| DC.type | 博碩士論文 | zh_TW |
| DC.type | thesis | en_US |
| DC.publisher | National Central University | en_US |