Chipola: A Chinese Podcast Lexical Database for capturing spoken language nuances and predicting behavioral data

Area of Science:

• Computational Linguistics
• Corpus Linguistics
• Psycholinguistics

Background:

• Spoken language analysis requires specialized lexical resources.
• Existing lexical databases may not fully represent the nuances of spoken Chinese.
• Podcasts offer a rich source for capturing authentic spoken language.

Purpose of the Study:

• To introduce Chipola, a novel Chinese Podcast Lexical Database.
• To analyze the characteristics of spoken Chinese vocabulary.
• To explore the applications of Chipola in language education and psychological research.

Main Methods:

• Development of a large-scale corpus from Chinese podcast transcripts.
• Extraction and analysis of lexical variables including frequency, context diversity, and part-of-speech.
• Evaluation of Chipola's performance against existing lexical databases.

Main Results:

• Chipola effectively captures core spoken Chinese vocabulary and linguistic features.
• The database demonstrates superior predictive power for third-party behavioral data compared to other resources.
• Chipola provides rich text-level information for simulating language input.

Conclusions:

• Chipola is an innovative and valuable resource for studying spoken Chinese.
• The database has significant implications for language education, offering pedagogical insights into language exposure.
• Chipola's utility extends to psychological research, particularly in understanding behavioral data.