Research On Personality Analysis And Privacy Protection Methods Based On Social Text

Cases of social engineering using people’s psychological weaknesses to attack emerge one after another and cause huge losses.Attackers develop a deeper understanding of the target person by discovering psychological attributes of the target with the help of data mining and artificial intelligence,so as to take a more accurate attack on the target.However,the current defense means are too passive and fail to grasp the core of social engineering.In psychological analysis,the psychological factors involved in people are relatively complex and numerous,and "personality" is a relatively comprehensive and stable psychological trait in the above research.Research shows that there is a clear correlation between social engineering attack and personality.In the era of information and big data,more and more people use social media to share their lives and views,making user’s text data open and accessible.It is the most effective and accurate method to analyze the personality information carried by text data by means of machine learning algorithm,and it is also one of the main channels for users’ personality privacy disclosure.Therefore,personality privacy protection based on text transformation is the most effective way to successfully block attacks.Firstly,based on the in-depth understanding and induction of personality analysis and prediction methods,aiming at the problem of revealing personality privacy from text data,in order to find the mapping relationship between text data and user personality,this thesis proposes a hierarchical hybrid model based on self-attention mechanism,namely HMAttn-ECBi L,to fully excavate deep semantic information horizontally and vertically.Multiple modules composed of convolutional neural network and Bi-directional long short-term memory encodes different types of personality representations in a hierarchical and partitioned manner,which pays attention to the contribution of different words in posts and different posts to personality information,and captures the dependencies between scattered posts.Moreover,the addition of word embedding module effectively makes up for the original semantics filtered by deep neural network.We verified the hybrid model on the My Personality dataset.The experimental results showed that the classification performance of the hybrid model exceeds the different model architectures and baseline models.In addition,protecting the personality privacy of the attacked object can effectively interfere with or deceive the attacker’s personality analysis and reduce the success rate of social engineering attacks.Therefore,this thesis proposes a text transformation method named Per Trans GAN,using generative adversarial networks(GANs)to protect the personality privacy hidden in text data.Based on the basic framework of the combination of GAN and reinforcement learning,we take the advanced text features and output probability value of the discriminator as the semantic signal and reward signal in the training process of text generator respectively.And,the loss function of the generator adds a penalty item to reduce the weight of words that contribute more to personality information in the real text,so as to hide the user’s personality privacy.In addition,the semantic and personality modules are designed to calculate the semantic similarity and personality distribution distance between the real text and the generated text as a part of the objective function.Experiments show that the model preserves users’ personality privacy as found in user data by transforming the text and preserving semantic similarity while blocking privacy theft by attackers.