A cheap and accurate delay-based IP Geolocation method using Machine Learning and Looking Glass

• DOI: 10.23919/ifipnetworking57963.2023.10186436
• Publication Date: 2023-07-24T17:38:31Z
• Authors: Allen Hong; Yahui Li; Han Zhang; Ming Wang; Changqing An; Jilong Wang
• Keywords: Geolocation

Predicting the geographical location of an IP host is a fundamental and valuable but long-standing challenge in field network research. Although delay-based methods have relatively high coverage low time consumption, currently this type method not accurate enough requires large number vantage points, making its cost high. In paper, we propose novel framework to make geolocation more cheap. Firstly, collect 373 Looking Glass with known addresses overcome problem by using them as points. Secondly, prediction coordinates machine learning algorithm regional information target IP. Finally, based on supplement missing values delay data improve accuracy successfully. Our experiment results validate feasibility improvement our method. Using method, average error 69.49 km for test set, which approximately 160 less than state-of-art work.