Search

Population Stereotype tells interaction designers just one-half of the complete story. It informs them only about the level of general consensus regarding each representation generated by different participants. It does not provide answers to those questions, which ask how the representation is to be achieved. Identification of different representation strategies adopted by different participants can reveal the rest of the story. In the presence of more than one or no strong contenders (population stereotype), adoption of the right representation strategy can be really beneficial. As most of the representational strategies are complementary to each other, the combination of different representational strategies can lead towards a more representative icon development.

The reduction of crop yield losses caused by pests is a major challenge to productive and sustainable food production for preventing food insecurity and emergencies, and for alleviating world food crisis. Multiple regression (MR) and artificial neural network (ANN) are two widely adopted modelling approaches for the prediction of crop pest risks, which are based on empirical statistics and artificial intelligence, respectively. Each of the two alternative approaches has its advantages and disadvantages. This study evaluates the two models from two aspects: their performances on pest risk prediction, and their methodological advantages and disadvantages. Two pest species are modelled using the two approaches as case studies, which are the melon thrip Thrips palmi Karny (T. palmi) and the diamondback moth Plutella xylostella (L.) (P. xylostella). Results show that ANN has higher prediction accuracy for both species. However, ANN has some methodological demerits compared to MR modelling.