In many Internet of Things (IoT) applications, knowing the device's location can be quite important for some reasons such as asset tracking and inventory management, geolocation services, safety and security, environmental monitoring, and proximity-based interactions. Mobile users often experience mobile services/applications within an indoor or outdoor environment. Operators and other service providers can offer more appropriate services to users if they predict their location. Several attempts have been made to categorize the location of users, but this paper proposes a methodology that increases the accuracy of the predicted values through Deep Neural Networks. Based on the proposed method, the accuracy of operator-side labels can be improved by comparing operator-side labeled datasets with real-world labeled drive-test collected datasets. The goal was to develop an accurate model to correct unassured labels of a dataset with more accurate labels based on three datasets collected with crowd-sourcing and drive-testing approaches. Also, the proposed method was compared with state-of-the-art learning algorithms in order to justify its superiority. The experimental results indicate that the F1-score metric may be as high as 98% in some datasets.