A hybrid of fuzzy c-means clustering and Latent Dirichlet Allocation for analysing philanthropic corporate social responsibility activities / Nik Siti Madihah Nik Mangsor

To date, philanthropic corporate social responsibility (PCSR) activities are ad-hoc in nature, where assistance is provided more to basic needs with very little attention to activities that can contribute to eradicating poverty. Based on previous related literatures, it is found that there is no proper categorization and documentation of CSR-related activities. Conventional clustering algorithms are able to extract only a single set of flat topics in which local and global topics are mingled and cannot be separated. Therefore, this research aims to develop hybrid fuzzy document clustering techniques to improve attribute cluster membership values of PCSR activities. This study has extended document clustering technique by integrating the traditional document clustering application with topic modeling approach. This integrated approach is able to produce precise results where it can help infer more coherent themes of PCSR activities. The analysis involved five-year data from the annual reports of 19 CSR-award winning companies in Malaysia where they were converted into a structured format, collated and summarized. Then, text pre-processing for data cleaning was performed followed by identification of the best Latent Dirichlet Allocation (LDA) topic modelling technique that was used to integrate document clustering.