因大數據時代的來臨,資料探勘的運用蓬勃發展,相較於較低結構的文本也成為眾多學者好奇的部份,在文字中所隱含的資訊儼然成了研究者探索的領域。本研究期望於農業教育之相關論文中挖掘出文字所隱藏的重要議題,透過分析結果給予國內當前農業教育的施行可參考方向。該研究利用VOSviewer和R語言等軟體工具,使用隱含狄利克雷分布(LDA)建立主題模型,找出國內外各年度中的熱門詞彙及主題內容,配合一致性分析決定主題數量。結果顯示,國內外文本主題皆著重於「飲食安全、營養與健康」、「氣候變遷的影響」、「都市農業」、「農業生產管理」和「永續農業」等議題,不同的是,國外文本中還探討了貧窮、性別地位和教育資源等議題。最後,能發現農業與環境、資源、人、食物、經濟、文化、生活和科技等部分的緊密關係,因此,農業的學習應該是多元且多面向的領域,應適時隨著環境而變動,不斷學習改變與突破,以及習得解決問題的能力,在農業上方能產生嶄新與創意來面對未知的改變。
Due to the advent of the Big Data era and the vigorous development of data mining, many scholars are curious about the text with the lower structure. Researchers explore the information implied in the text. The purpose of this research is to dig out important topics hidden in the text from papers related to agriculture and education, and to give reference directions to the current domestic agricultural education. The research uses software tools such as R and LDA to build a topic model, finds out popular termsand topic contentsin every year, and determines the number of topics with coherence. The results show that both domestic and foreign topics are focused on such as “food safety, nutrition and health”, “impact of climatechange”, “urban agriculture”, “agricultural production management” and “sustainable agriculture”. The difference is that foreign texts discuss issues such as poverty, gender status, and educational resources. Finally, it is possible to discover the close relationship between agriculture and the environment, people, resources, food, economy, culture, life, and technology. Therefore, agricultural learning should be a diversified and multi-faceted field, which should change in time with the environment. Constantly learning changes and breakthroughs, as well as acquiring the ability to solve problems, can generate new and creative ideas to face unknown changes in agriculture.