In this directory, we present the supplemental material for our paper entitled "On Automated Trust Computation in IoT with Multiple Attributes and Subjective Logic". The following files are included in this directory:
- IssuesWithPreviousNormalizationMethods.ipynb: We explain attribute normalization methods that are previously proposed in seminal works of MADM [1] [2]. We discuss the issues that we detect in using these methods in the context of trust computations for IoT. We explain those isues in detail and present sample scenarios where they can be observed.
- weight_approximation.ipynb: In this file, we present surrogate weights that we obtain by applying different weight approximation methods, namely Rank Order Centroid (ROC), Rank Sum (RS), and Rank Reciprocal (RR).
- converting_scores_to_intervals.ipynb: In this file, we present a method for converting trust scores into an interval that is an alternative to the method we present in the paper (in the following section: Converting Trust Scores into Opinion Triplets).
- sample trust computations.pdf: We demonstrate our trust computation method in this file, using the trust measurement example given in the paper (in Section IV) and a sample trust network.
[1] Hwang, C. L., & Yoon, K. (1981). Methods for multiple attribute decision making. In Multiple attribute decision making (pp. 58-191). Springer, Berlin, Heidelberg.
[2] Yoon, K. P., & Hwang, C. L. (1995). Multiple attribute decision making: an introduction (Vol. 104). Sage publications.