One issue investigated in this paper is the influence of different types of fuzzy measures on classification accuracy obtained from the Choquet integral based combination of feedforward neural networks. The ordinary fuzzy measure, the lamda-fuzzy measure, the 2-additive fuzzy measure, and the fuzzy measure depending only on set cardinalities are used in the study. The second issue investigated in this paper is the effectiveness of the half&half sampling approach in creating accurate neural network committees fused through the Choquet integral. The classification performance obtained from the Choquet integral based neural networks committee is compared with the performance of the weighted averaging combination scheme, the majority voting scheme, and the performance obtained from the best single neural network.