Neural Networks (NNs) are non-linear models and are widely used to model complex relationships, thus being well suited to formulate the product design process for matching design form elements to consumers’ affective preferences. In this paper, we construct 36 deep NN models, using one to four hidden layers with three different dropout ratios and three widely used rules for determining the number of neurons in the hidden layer(s). As a result of extensive experiments, the NN model using one hidden layer with 140 hidden neurons has the highest predicting accuracy rate (80%) and is used to help product designers determine the optimal form combination for new fragrance bottle design.

In the traditional product design field, most studies focus on commercial products (e.g. furniture, vehicle, consumer electronics, and so on), while few studies pay attention on the craft art. As such, this paper conducts a consumer-oriented experiment on characteristic toys to determine how to best meet consumers' preferences for new product design. The characteristic toy is a good illustration to be an application of art and design, because it is not only one type of craft art, but also a commercial product. The characteristic toy has two features simultaneously of craft art and product design. In this study, the new approach proposed consists of two main methods, i.e. the quantification theory type I (QTTI) and the technique for order preference by similarity to ideal solution (TOPSIS). There are three QTTI models used to build a design decision support database. According to the design decision support database, a QTTI-based TOPSIS model is used to help product designers to obtain the optimal design alternatives that best meet consumers' preferences.