Trend feature-based clustering for research funding time series data

This paper presents an efficient computational method for time series clustering and application concerning research funding of universities directly under Minster of Education of People Republic of China. Presented approach was based on extraction of trend features with Haar wavelet decomposition from time series data and their use in feature-based agglomerative hierarchical clustering of monthly measured research funding income data. This method could be implemented for users who desire to manage research funding of all universities.