| Abstract [eng] |
The dissertation analyses the determination of fraud in the financial statements: the conception of fraud was analysed from different perspectives, the components of fraud conception and the conditions influencing the fraud were distinguished. The conception of fraud determination in the financial statements was identified as a seperate type of fraud and in order to define it better the qualification of frauds in the financial statements was created. The thesis analysed the determination process of fraud in the financial statements and systematised the actions of fraud determination. The results of the carried out research of tendencies of fraud in financial statements in Lithuania revealed the need of model creation. Having examined the advantages and disadvantages of fraud determination models in financial statements, analysed in the scientific literature and taking them into account, a model of fraud in the financial statements, including the determination actions used in accounting, audit, economics, law and informatics, was created. The adapted SOM neural networks and the created corporate database when used together can attribute the investigated company to those which are characteristic to fraud or not. In order to reach the model efficiency, the expert assesment of it was caried out by: FCIS, FSCL, STI and auditors and taking into account the assesment results, the created model was slightly improved. In order to find out the practical application of the model, it was empirically tested in the example of a Lithuanian company. |
