應用類神經網路模擬產品功能屬性組合搭配之研究

全球化的發展趨勢，使得人口族群與生活型態不斷地在轉變，這讓產品的定位與消費者的需求偏好，愈來愈難以掌握，產品的設計亦變得多樣且複雜。因此，本研究針對現有吹風機的功能與特性，利用電腦資料運算的優勢，透過類神經網路進行解析，搜尋滿足消費者的產品屬性組合，藉此推估及預測設計的產品受消費者的滿意程度，有效地將顧客需求聲音轉化成實際產品的特色，提供設計人員迅速且正確的掌握消費者資訊。研究結果顯示，經由推演模擬28組的滿意度問卷調查，可推論出其餘1847組的滿意結果，設計人員即可有效地選用適切的功能特性之組合，因應市場同質產品重複修正與淘汰的壓力，以及改善必須仰賴設計經驗的傳承或依靠設計師的經驗脈絡來完成產品的設計。

關鍵字

吹風機；產品設計；類神經網路；人工智慧

並列摘要

The development of trends of globalization and global village accelerates tangible and intangible communication, which transforms populations, races, and life styles continuously and renders the positioning of products and the demands and preferences of consumers harder to grasp; product design has also become diverse and complex accordingly. Therefore, it is to perform research and analysis mainly on the hair dryer's attributes such as functionalities, size, weight, performance, and safety etc. through parallel information processing technologies. This system will simulate the satisfaction rating for the product characteristics through neural networks which implement neuron network and genetic algorithms, to search for the product attribute combination that satisfies consumers, and to estimate and forecast whether consumers will be satisfied with the designed products. In doing so, the voices for the consumer demand can be effectively transformed into special features of an actual product and consumer information can be obtained quickly and correctly. The research results show about combination of product function features: the simulation using neural network can deduce the rest of the 1847 sets of the satisfaction results from 28 sets of satisfaction survey, which can be effectively used by designers to choose the appropriate combination of functional features.