Predictive Wi-Fi Network Selection: Enhancing User Experience through a Machine Learning-based Approach

With the increasing use of Wi-Fi in our devices at home and in offices, etc. it’s imperative to connect to an AP which provides seamless and smooth performance. This research paper addresses the need for an efficient solution to predict and select the most suitable Access Point, thus enhancing the user’s quality of application usage. We have worked on a Machine Learning model which takes care of different parameters such as signal strength, bandwidth, distance, frequency, traffic parameters and previous connections with Access Point and predicts the best Access Point user should connect to. The methodology described in this paper has been developed based on the Google scoring algorithm, which we referred as a foundation for our approach. The integration of the Google scoring algorithm empowers our model to further optimize Access Point selection, ensuring an enhanced user experience while seamlessly switching between networks, thus solving the Sticky band or Sticky Client problem in Wi-Fi. The research which has been proposed by this paper could potentially revolutionize the way transitions are made between networks to make the process smooth and latency free.