1 Introduction

Translation quality assessment is important in multiple fields, but existing research often uses descriptive methods and lacks quantitative analysis. This article quantitatively studies relevant literature using Python and word vector clustering techniques, aiming to comprehensively display the research status and trends, and propose improvement suggestions, providing reference for interdisciplinary research.

2 Literature Review

Explored in depth the current status of translation quality evaluation for both manual and machine translations. In terms of manual translation evaluation, various evaluation models have been introduced, such as the House model, Reiss model, and Williams model, which are based on systemic functional linguistics, discourse type principles, and argument structure analysis, respectively. Meanwhile, domestic scholars such as Si Xianzhu and He Sanning have also explored evaluation theory. The application of Professional Language Assessment (LAPP) in the ESP field emphasizes the importance of language knowledge and specialized content knowledge. The differences in research paradigms between domestic and foreign studies are particularly evident in the evaluation of legal translation. Domestic scholars tend to focus more on translation norms and terminology translation quality, while foreign scholars tend to use scale scoring and pre selection evaluation. In addition, the development of computer-aided evaluation models, such as semantic feature-based machine scoring models and automatic scoring systems, was also mentioned.

3 Experimental Design

The research subjects are the Web of Science core dataset and CNKI indexed literature, with keywords of "translation quality assessment" and "translation quality", covering the fields of humanities, social sciences, and natural sciences. A total of 1575 articles were collected, and Python automated scripts and word vector clustering were used to process the data. Data processing includes keyword extraction, core keyword filtering, word vector expression, and feature word clustering, using tools such as SpaCy, jieba, Word2Vec, PCA, and K-Means.

4 Research Results

The development process of translation quality assessment research is shown through annual statistics and trend charts, which can be divided into three stages: embryonic stage, slow development stage, and fluctuating upward stage. Research shows that the overall number of literature in the field of translation quality assessment is on the rise. Keyword analysis reveals the differences in research hotspots between CNKI and WOS databases, with CNKI focusing on theoretical research and WOS focusing on automatic evaluation of machine translation. Influential scholars and highly cited literature analysis indicate that scholars in the field of computer science play an important role in evaluating translation quality, while scholars in the field of linguistics have relatively less influence. There are differences in research perspectives and emphasis between Chinese and English literature, with CNKI literature focusing more on theory and WOS literature focusing more on technology.

5 Problems and Countermeasures

Although there has been progress in the field of translation quality assessment, there are still problems such as slow research development, isolation in humanities and sciences research, and a disconnect between research and practice. The research began in the 1980s with a low number of publications and has shown a downward trend in recent years. The research in the fields of computer science and translation is independent of each other, while the research in the ESP field is scattered. The existing theoretical models are disconnected from practice, and there is a gap between technology and practice. To address these issues, it is recommended to establish a three-dimensional and diverse academic promotion system, strengthen government research fund investment, increase academic conferences and evaluations, and establish a data crowdsourcing platform. Adhering to the concept of interdisciplinary research, we will jointly build a channel for cooperation between arts and sciences, and create a research model that promotes the integration of arts and sciences. Improve the practice oriented translation evaluation path, smooth the flow path between theory and practice, and establish a linkage mechanism between technology and practice. Build a human-machine collaborative quality evaluation model, strengthen the platform based research and development of automatic evaluation technology, and improve the technical literacy of translation evaluators. This article systematically reviews the development of the field of translation quality assessment through quantitative analysis, and proposes existing problems and countermeasures.