身處資訊爆炸的時代，資訊內容及其可供選擇的數量愈來愈多，相形之下人們對於從大量搜尋結果中進行決策就愈發困難。推薦系統係因應如此之需求而產生，它不只降低人們搜尋的成本，更協助人們發掘潛在感興趣的事物。
隨著網際網路的盛行及資訊科技的快速發展，近幾年愈來愈多的企業及民眾透過網際網路的管道，例如：線上人力銀行網站、社群網路等徵才及求職。但綜觀人力銀行徵才系統，雖具有龐大的求職庫，但對使用者而言，往往要在海量的資料中過濾出符合期望的工作實屬不易。
有鑑於此，本研究嘗試開發一套以「基於比較」應用交談式推薦技術之虛擬人力銀行網站系統，該技術將依據工作職缺特徵及求職者偏好進行相關資料分析、匹配及推薦，藉由開發以使用者「偏好回饋」模式及「基於比較」之架構與機制，並導入「多樣性推薦」演算法，實證研究交談式推薦機制的績效。
It is increasingly difficult to make effective use of Internet information, given the rapid growth in data volume and diversity. That is how the recommendation systems come into play in this field. Recommendation systems not only for helping people to choose products or services, but also reducing them the time spent in the searching in electronic commerce.
With the rapid development of the Internet and information technology, many businesses and job seekers through the Internet such as online job banks, social networks as channel for jobs in the recent years. Although people looking for jobs on the Internet could be a great way of finding new and interesting opportunities, it is difficult to find suitable job information for them in a large amount of occupation data.
In this research, there was a comparison-based recommendation approach developed which avoids engaging the user in a deep dialogue in favour of a more casual conversational strategy. This work and experiments were carried out using the dataset from the online job bank domain and presented a novel web-based conversational recommendation system which attempts to use the preference feedback as a means of revising the query and embed a new retrieval algorithm called similarity-preserving to increase in diversity. In summary, this research has demonstrated promising results for the comparison-based recommendation approach, in terms of its effectiveness and efficiency characteristics.
