The linguistic Multi-Criteria Group Decision-Making (MCGDM) problem involves various types of uncertainties. To deal with this problem, a new linguistic MCGDM method combining cloud model and evidence theory is thus proposed. Cloud model is firstly used to handle the fuzziness and randomness of the linguistic concept, by taking both the average level and fluctuation degree of the linguistic concept into consideration. Hence, a method is presented to transform linguistic variables into clouds, and then an asymmetrical weighted synthetic cloud is proposed to aggregate the clouds of decision makers on each criterion. Moreover, evidence theory is used to handle the imprecision and incompleteness of the group assessment, with the belief degree and the ignorance degree. Hence, the conversion from the cloud to the belief degree is investigated, and then the evidential reasoning algorithm is adopted to aggregate the criteria values. Finally, the average utility is applied to rank the alternatives. A numerical example, which is given to confirm the validity and feasibility, also shows that the proposed method can take advantage of cloud model and evidence theory to efficiently deal with the uncertainties caused by both the linguistic concept and group assessment.

We describe here development of a multichannel high-Tc SQUID magnetometer system for measurement of cardiac magnetic fields, aiming at future application of diagnosis of heart diseases. Two types of direct-coupled SQUID magnetometers were fabricated and used: single pickup coil magnetometer having flux dams to suppress the shielding current that would induce flux penetration and the consequent low-frequency noise, and double pickup coil magnetometer having no grain boundary junctions and flux dams on the pickup coil. The superconducting film of both the magnetometers had holes and slots, leaving 5 µm-wide strip lines, to suppress trapping and penetration of magnetic flux vortices in environmental fields. We studied different schemes of active shielding to reinforce the efficiency of field-attenuation of magnetically shielded room (MSR). A feedback-type compensation using a normal detection coil wound around the wall of MSR and a selective cancellation of 50 Hz noise by means of adaptive filter were developed. Such combination of passive and active shielding, based on the use of simple MSR, would be suitable in a practical low-cost magnetometer system for clinical MCG examination. We fabricated a liquid nitrogen cryostat that could contain up to 20 magnetometer-capsules at 4 cm separation in a flat bottom, with a distance of 16 mm between the air and liquid nitrogen. The cryostat was set in a gantry, which had rotational, vertical and horizontal freedoms of movement, in a moderate-shielding MSR that was combined with the developed active shielding. Measurements of MCG were performed for normal subject using eight magnetometers operating simultaneously.

An algorithm based on the wavelet transform (WT) was developed to analyze the QRS complex in a three-dimensional magnetocardiogram (3-D MCG) recorded from 3 normal subjects and 1 patient with anterior myocardial infarction (MI). By using a wavelet equivalent filter constructed with the WT algorithm, the high frequency components of the QRS complex related to the late fields (LF) were detected for the patient with anterior MI at different scale. We quantified the high frequency components of the QRS complex by calculating root-mean-square (RMS) value at different scale. The LF mainly existed in the frequency band of about 35.5 to 110.5 Hz with the amplitude of about 0.1 to 0.4 pT for Bx, By, and Bz components. In order to discuss the activities of the heart between the normal subject and the patient with anterior MI, we have also evaluated the spatial energy distribution (SED) of the QRS complex by displaying isoenergy contour maps at different scale. Being different from the normal subject, the patient with anterior MI represented different the pattern of the SED in various frequency band for the ST segment of the QRS complex of Bx, By, and Bz components. It is efficient to use the WT algorithm for analyzing the QRS complex in the 3-D MCG.

The aim of this study was to quantify the effects of inhomogeneities on magnetocardiography (MCG) forward solutions. It can serve to guide the selection of inhomogeneities to include in any geometric model used to compute magnetocardiographics fields. A numerical model of a human torso was used which construction included geometry for major anatomical structures such as subcutaneous fat, skeletal muscle, lungs, major arteries and veins, and the bones. Simulations were done with a single current dipole placed at different sites of heart. The boundary element method (BEM) was utilized for numerical treatment of magnetic field calculations. Comparisons of the effects of different conductivity on MCG forward solution followed one of two basic schemes: 1) consider the difference between the magnetic fields of the homogeneous torso model and the same model with one inhomogeneity of a single organ or tissue added; 2) consider the difference between the magnetic fields of the full inhomogeneous model and the same model with one inhomogeneity of individual organ or tissue removed. When single inhomogeneities were added to an otherwise homogeneous model, the skeletal muscle, the right lung, the both lungs and the left lung had larger average effects (15.9, 15.1, 14.9, 14.4% relative error (RE), respectively) than the other inhomogeneities tested. When single inhomogeneities were removed from an otherwise full inhomogeneneous model, the both lungs, the left lung, and the skeletal muscle and the right lung had larger effects (17.3, 14.9, 14.3, 14.2% relative error (RE) respectively) than other inhomogeneities tested. The results of this study suggested that accurate representation of tissue inhomogeneity has a significant effect on the accuracy of the MCG forward solution. Our results showed that the inclusion of the boundaries also had effects on the topology of the magnetic fields and on the MCG inverse solution accuracy.

Feedback of class memberships is incorporated into multimodal pattern classifiers and their unsupervised learning algorithm is presented. Classification decision at low levels is revised by the feedback information which also enables the reconstruction of patterns at low levels. The effects of the feedback are examined for the McGurk effect by using a simple model.