The auditing system is one of the most important systems for supervising and maintaining state governance in modern countries. The duties of auditing institutions include not only supervising budget execution, reviewing financial revenues and expenditures, verifying final accounts, assessing financial efficiency, inspecting financial losses, and verifying financial responsibility. It is also possible to promote national transparency, ensure accountability, improve performance, and fight corruption through the exercise of authority to improve the country's good governance.
The data application workshop cooperated with the National Audit Office (NAO). After the discussion, the workshop decided to assist the NAO to do the text mining in
accounting voucher summaries to find out the main keywords and help future audit work of the NAO. The NAO has a large amount of data of accounting voucher summaries information, with a number of hundreds of millions of copies. Such huge amounts of data cannot be surveyed by traditional manual methods. Therefore, in order to improve the effective supervision and inspection of auditing units, the use of machine-based text mining has become a better solution. The outcome of the cooperation of data application workshop and the NAO is use text mining technique to find out important keywords, conduct basic word segment analysis, establish a keyword list, and then enter voucher data by months according to the internal voucher database of the NAO. With the keyword list, the data table is established to construct the preliminary search and result presentation system. After the user inputs the keyword query, the relationship tree is generated according to the voucher association table, and the user can generate tree diagram of the desired query by the appearance frequency of the keyword and the keyword relationship level setting conditions. The results are shown in the following figure:
In addition to the generation of preliminary analysis results, the following recommendations have been found in the data application workshop. First of all, due to the large number of accounting vouchers in the NAO, the machine-based keyword association analysis is limited by the software and hardware computing ability. The system needs to be developed in the future for the use of all auditors. It may be necessary to further consider the relevant software and hardware upgrade planning. Secondly, the case of the cooperation between the data application workshop and the NAO has become a model of cooperation for the data application workshop, which can be used as an example for subsequent workshops.