Leveraging Data Trust worthiness in Mobile Crowd-Sensing using GAN

Smartphones have become very commonplace. Nearly, 6.94 billion people out of 8 billion people on the earth, own a smartphone. This high rate of adoption implies that cell phones might not be a luxury for many people around the world, but rather a vital tool. All the data collected from these smartphones holds the potential for significant advancements, and ensuring that the collected data is legit and integral is essential. This paper introduces an innovative algorithmic approach that distinguishes between genuine and malicious users on crowdsensing platforms. The model is initially trained with the input dataset and determines the legitimacy of the newly collected data. The combination of Generative Adversarial Networks (GAN) and Random Forest Classifiers in our model leverages data trustworthiness in mobile crowd-sourcing. The model employs a generator to produce new data, which is then utilized to augment its training process. By generating synthetic data, the model can expose itself to a broader range of scenarios and variations present in real-world data. This augmentation process helps the model to learn more effectively from the available training data, potentially leading to better performance on unseen examples during inference. Through this comprehensive approach to data trustworthiness, this model offers a solution for refining the crowd-sensing data.