Abductive reasoning networks

Modeling Obesity Using Abductive Networks

“…This makes AIM much easier to use and considerably reduces the learning/development time and effort. AIM advantages over back-propagation neural networks in forecasting the daily minimum temper- ature were demonstrated (27 ), and large improvements in training speed were reported (24 ). It should be mentioned, however, that improvements are becoming available which attempt to alleviate some of the problems associated with older classical neural network paradigms.…”

“…Abductive networks (24 ) combine the advantages of the neural network approach with those of advanced statistical methods. The AIM tool is claimed to be faster, more accurate, more automated, and easier to use than alternative neural network and statistical techniques for a large class of problems (24 ). While the processing elements in neural networks are restricted by the neuron analogy, AIM builds networks of various types of more powerful numerical functional elements based on prediction performance.…”

“…Such variations have led to newer techniques combining good features from various methods. An example of such hybrids is the AIM (Abductory Induction Mechanism) 1 abductive network tool, which draws on statistical and multiple regression analysis methods as well as neural networks, resulting in faster and more automated model synthesis (24). The development of this approach for machine learning through self-organization has followed the track of the group method of data handling (GMDH) algorithm (25 ) and the closely related adaptive learning network (ALN) technique (26 ).…”

Porosity prediction using the group method of data handling

“…The predicted output layer of the GMDH network can be represented as a polynomial function of some or all of the inputs. Campbell and Johnson (1993), Montgomery and Drake (1990), and Shastri et al (1998) find GMDH prediction to be significantly more robust and accurate than those of the ANN. GMDH has been used in the classification, in determining which inputs are more important to the modeled system, and in predicting the outputs of complex systems (Barron et al, 1984).…”

Modeling and forecasting electric daily peak loads using abductive networks

“…To overcome such limitations, we propose using the alternative machine learning technique of abductive networks [19] for daily peak load forecasting. We have previously used this approach to model and forecast next day's hourly load profile [20], monthly domestic electric energy consumption [21], and the minimum and maximum daily temperatures [22,23].…”