基于屬性重要性的手勢改善方向決策算法

婁澤華　殷繼彬

摘要：為了更有效地改善手勢以延長其生命周期，采用量化指標指導手勢改善方向的決策。基于屬性重要度，給出了復雜人因條件下手勢改善方向的決策算法。根據(jù)模糊層次分析法計算指定手勢每個屬性的全局權(quán)重，并根據(jù)用戶反饋計算指定手勢每個屬性的局部權(quán)重，全局權(quán)重與局部權(quán)重調(diào)和得到綜合權(quán)值向量。依據(jù)用戶對指定手勢的綜合印象，將各屬性評分分為兩類分別進行處理。根據(jù)各屬性的正向與負向影響力，得到權(quán)值向量中各屬性權(quán)值分布。對各屬性評分分別計算標準化評分偏置，各屬性改善需求程度排序通過權(quán)值分布與標準化評分偏置進行計算。實驗結(jié)果表明，基于該算法比基于問卷調(diào)查決策制定的手勢，支持率平均提高了25%，從而得出結(jié)論：手勢優(yōu)化過程中各屬性的權(quán)值排序是穩(wěn)定的。

關(guān)鍵詞：屬性重要性；手勢優(yōu)化；決策算法；人機交互；權(quán)值分布

DOIDOI：10.11907/rjdk.181208

中圖分類號：TP301.6

文獻標識碼：A文章編號文章編號：16727800（2018）009001309

英文標題Attribute Importancebased Decision Algorithm for Gesture Direction Improvement

--副標題

英文作者LOU Zehua， YIN Jibin

英文作者單位（Department of Information Engineering and Automation，Kunming University of Science and Technology，Kunming 650500，China）

英文摘要Abstract： In order to improve the gestures more effectively to prolong their life cycle，a quantitative indicator is used to guide the gesture to improve the direction of the decision.Based on attribute importance，an algorithm is proposed to make decision on gesture improvement direction under complex human condition. The global weight for each attribute of the specified gesture is calculated according to the FAHP （Fuzzy Analytic Hierarchy Process） algorithm.The local weight for each attribute of the specified gesture is calculated according to the user feedback.The global weights and the local weights are harmonized into comprehensive weights vector.According to the user's comprehensive impression on the specified gestures，each attribute's score is divided into two classes and processed separately.According to the positive and negative influence degree of each attribute，the weight distribution of each attribute in the weight vector is obtained.Each score of the attributes is calculated by the standardized score bias respectively.The rank of improvement necessity on each attribute is calculated by the weight distribution and the standardized score bias.Experiments show that the support rate of improved gestures achieved through decisionmaking based on this algorithm is 25% higher on average than based on questionnaire.The conclusion is that the weight ranking of each attribute is stable during the process of gesture optimization.

英文關(guān)鍵詞Key Words：attribute importance；gesture optimization；decisionmaking algorithm；humancomputer interaction；weight distribution

0引言

手勢是以人因（人為因素）為導向的，其優(yōu)劣程度并沒有精確的評判標準。當前人機交互類產(chǎn)品最具代表性的客觀和主觀評價標準分別是產(chǎn)品可用性與用戶體驗（通常縮寫為UX）。很多從業(yè)者從以下幾方面考慮產(chǎn)品可用性：靈活性[1]、可學習性、可記憶性與安全性。產(chǎn)品可用性在ISO9241 Ergonomics of Human System Interaction標準（1998年第11部分）[2]中被定義為：指定用戶在指定情境中使用指定產(chǎn)品（服務或環(huán)境）實現(xiàn)指定目標時的有效性、效率與滿意程度。……