원문정보
초록
영어
The fast-developing Chinese higher education and research institutions are faced with more and more serious challenges in their financial regulation. Traditional accounting approaches may fail to distill useful decision-making suggestions confronted with the multi-species and huge-quantity financial “big data”. To reveal valuable information, this research focuses on the pattern recognition of the funding flows in 76 universities under Chinese Ministry of Education. Given the trend feature of the data series detected by the Mann-Kendall Non-Parameter Ranking Test, the low-frequency parts of the funding flows are distilled with wavelet transform to represent their basic features. Results show a clear hierarchy structure in the investments and expenses of universities according to their “titles” and advantage disciplines. Specifically, the comprehensive universities fall to different categories according to their “titles”, while the professional universities are classified according to their disciplines. The scientific and technical focused universities show larger variance among different categories than other specific discipline focused universities.
목차
1. Introduction
2. Methodology
2.1. Trend Test
2.2. Data Feature Extraction
2.3. Similarity Measurement
2.3. Clustering
2.4 Evaluation Indicators
3. Experiment
3.1 Data Description
3.2 Data Trend Test
3.3 Clustering Based on Data Features Extraction
4. Results
4.1. Clustering Results
4.2. Evaluation Results
5. Discussion
6. Conclusion
Acknowledgment
References