自動化語言環境分析：應用於嬰幼兒之效度綜論

Automated Language Environment Analysis: A Comprehensive Review of Validity in Infant and Young Child Applications

家庭語言環境在幼兒語言和認知能力發展上扮演重要角色。過去語言環境的研究多透過人工蒐集與分析語言資料，耗時耗力，難以即時應用，但隨著科技進步，語言環境分析系統（the Language ENvironment Analysis System, LENA）被開發，此為一種結合穿戴錄音器和自動語音分析軟體的研究工具，能長時間蒐集自然環境中的語音，並提供分析結果。隨著LENA被廣泛應用於語言相關研究，針對其效度進行探究的研究也日益增多。本研究透過文獻分析，探討LENA在單一語言（如英語、華語、西班牙語等）環境之效度分析樣貌，期作為未來效度分析之參考，幫助語言研究更有效地應用LENA，並促進語言研究與人工智慧工具結合，擴展相關研究之範疇。

The home language environment plays a crucial role in the development of children’s language and cognitive abilities. Studies on language environments have primarily relied on the manual collection and analyses of linguistic data. This process is time consuming and labor-intensive, consequently limiting the real-time application of such data. Technological advancements have facilitated the development of an automated speech analysis tool known as the Language ENvironment Analysis (LENA) system. The LENA system is a research tool that integrates a wearable recording device with automatic speech analysis software to enable the collection of speech data in natural environments over extended periods of time and provide analysis results. With the LENA system becoming widely applied in language-related research, the number of studies evaluating its validity has also increased. Using a literature review, this study explored validity analyses conducted for the LENA system in various linguistic environments, such as English, Chinese, and Spanish. The findings of this review may not only serve as a reference for future studies on the validity of the LENA system but also aid in the more effective application of this system in language research. Accordingly, the integration of language research and artificial intelligence techniques can be promoted to expand the scope of related studies.