Analysis of Transportability of Causal Effects in Wireless Sensor Networks
AbstractIn the empirical sciences, experiments are carried out in the laboratory so that the results may be applied in other locations where the conditions are the same as in the laboratory. Hence, it should be possible to establish certain causal statements in the laboratory, which can be transported to real-world environment. In the present paper, experiments were carried out on wireless sensor networks in the laboratory and the results were transported to environments where experimentation is impossible. The tool Tetrad was used to arrive at the causal directed acyclic graph for the wireless sensor network in the laboratory. The concepts of selection diagram, the principle of transportability and the condition for failure of transportability are discussed. The applicability of transportation algorithm for typical environment and also the failure of the same are analyzed. Simulations are carried out using the simCausal package in R programming and the validity of the probability expressions transported by the transportation algorithm are shown to be correct.