Search

Privacy statements are legal documents designed to protect the interests of companies and individuals, most commonly in web services and are widely adopted way of informing users how data will be collected, stored, shared. However, the average Internet user either does not read these statements or it takes considerable amount of time and yet it's difficult to fully understand the statement content and meaning. On the other hand, the length and complexity of the privacy statements are an obstacle for regulators to verify that the requirements are applied. Utilization of natural language processing allows document's processing to be automated to some extent and thus to reduce the gap in understanding the content by the end users. A model for automated analyses of privacy policy documents based on natural language processing and machine learning is suggested in the paper. The model comprises a pre-processing stage for data retrieval and fusion of information from different sources as well as data processing stage that utilizes neural network for data classification. The neural network architecture and several groups of model parameters have been experimentally evaluated using a dataset created using five information sources of both publicly available annotated privacy policy documents and self-annotated documents. [ABSTRACT FROM AUTHOR]