Inference by aggregation of evidence with applications to fuzzy probabilities

Multidimensional Conceptual Spaces for Summarization

“…If a crisp value is desired, defuzzification must be done either before the conditional probabilities are computed, or after. In the latter case, fuzzy arithmetic rules, with the additional constraints that the result must be a fuzzy set in [0,1] are used to compute the conditional probabilities as fuzzy sets [8]. In the experiments carried out for this study cardinalities are first defuzzified and then used to compute a unique/crisp value for the matrix entry.…”

Random weighting estimation for fusion of multi-dimensional position data

“…Although there have been a significant amount of research efforts reported during the past twenty years, fusion of multi-sensor data is still a challenging problem [8,9,17]. Currently, the commonly used methods such as Kalman filter [1,2,15,16], Bayesian reasoning [3,12,19] and fuzzy logic theory [13,14,18] suffer from their own limitations in achieving optimal fusion. Such limitations include the dependence on a conditional probability distribution or fuzzy membership function, the unacceptable fusion result when observational evidences highly conflict with each other, the low real-time performance due to the use of too many state variables, and the low efficiency for fusion of multi-sensor information [8,10,11].…”

A meaning based information theory – informalogical space: Basic concepts and convergence of information sequences

“…Instead of computing with numbers, much research has focused on tackling the problem of computing with words, meaning, inference and reasoning based on meaning (see [7,12,14,28,30,35,40,41,[49][50][51]). …”