智能手机可用性综合评价模型构建

发布时间：2018-06-13 00:41

本文选题：手机可用性 + 可用性指标　；参考：《浙江理工大学》2014年硕士论文

【摘要】：近年来，产品的可用性愈来愈受到人们的重视，手机在人们的生活日益普及，手机的可用性也渐渐得到了用户的重视，但在手机开发过程中，往往存在追求技术创新而忽略人的因素的现象，从而导致目前许多手机产品都存在不同程度的可用性问题。对于手机可用性的指标，在以往研究中也存在着单一化的局限，（1）仅测量主观指标；（2）仅测量客观指标;（3）客观指标的测量只有结果指标。同时由于主客观绩效出现的分离，在对可用性差异的比较上，往往难以进行。因此本研究试图使用多种方法多种手段来对手机可用性进行测量，从负荷的角度，使用过程性的可用性指标（生理指标）来测量手机的可用性，同时综合主观指标和不同类型的客观指标，建立可用性的指标体系，并通过算法来对可用性的指标体系进行相互验证和预测，通过指标之间的匹配,来解决有效的进行不同手机之间的可用性差异比较的问题。本文中研究一确定了指标体系中的维度，（1）实验一为评价体系中的主观维度的测量编制了手机可用性问卷，使用223名被试进行了项目的初测，通过因素分析确定了问卷的七个维度，最终确定的问卷一共40个项目，具有良好的信度与效度。（2）实验二验证了肌电指标作为客观过程指标的有效性，通过操作同一手机，控制任务难度水平，验证肌电对不同负荷水平的区分，从而验证肌电作为可用性指标的有效性。采用20个被试的有效数据，结果得到在两种任务难度水平（简单/困难）上的肌电平均振幅（AEMG），中位频率（MF），平均功率频率（MPF）的多数任务上均有显著的差异，在复杂任务上肌肉的负荷较大。研究二基于BP神经网络算法构建了可用性综合评价模型（3）实验三中测量了多个维度的手机可用性综合指标，为下一步模型的构建建立了基础。采用组间设计，两组被试分别使用HTC手机和三星手机完成操作任务，记录被试的任务完成时间和操作过程中的肌电数据，任务完成后填写手机可用性问卷。对实验三的结果进行了差异性比较，在问卷分数上，两个手机没有显著差异（t=-0,569.p=0.572）。在任务完成时间上，两个手机在任务一、任务八、任务十二和任务十四存在显著性差异(t=-2.420*,p=0.020;t=-3.413**,p=0.001;t=2.538*,p=0.015; t=-3.338**,p=0.002),在AEMG上，两个手机没有显著性差异。对于MF，除了任务8，任务14，在其他任务上两个手机均出现了显著差异（t=2.455，，p=0.018；t=2.619，p=0.012；t=2.696，p=0.010；t=2.623，p=0.012；t=3.436，p=0.001；t=3.526，p=0.001；t=-2.220，p=0.031；t=7.397，p=0.000；t=9.077，p=0.000；t=5.945，p=0.000；t=4.694，p=0.000；t=8.481，p=0.279），具有显著性差异的这些任务中，任务7上HTC比三星更加省力，其余任务均是三星手机的表现更好。对于MPF，两个手机在任务1，2，3，4，5，7，8，10，12，13，14中均出现显著性差异（t=4.450,p=0.000;t=6.021,p=0.000;t=2.942,p=0.005;t=6.099,p=0.000;t=3.187,p=0.003;t=8.475,p=0.000;t=4.335,p=0.000;t=5.046,p=0.000;t=7.303,p=0.000;t=4.070,p=0.000）,在这些具有显著性差异的任务中，均为三星手机更为省力。（4）在实验三中通过任务的对比不能区分两个手机的可用性差异，因此在下一步进行了建模，基于BP神经网络算法，通过非线性的映射，尽可能地利用了数据，建立了预测准确性在73.3%的可用性评价模型。获得了数据在23个维度上的权重，通过计算更加准确客观地对两个手机的可用性差异进行了对比。本文综合了多种可用性评价指标对手机可用性进行评价，克服了以往研究中评价指标单一的局限。并通过算法建立了评价模型，从模型的输出对不同手机的可用性差异进行了客观的比较。
[Abstract]:In recent years, the availability of products has been paid more and more attention. Mobile phones are becoming more and more popular in people's life, and the availability of mobile phones has gradually been paid attention by users. However, in the process of mobile phone development, there are often the phenomenon of pursuing technological innovation and ignoring human factors, which leads to the existence of many mobile products at present to different degrees. Availability problem. There is also a single limitation in the previous research on the availability of mobile phones, (1) only the subjective indicators are measured only; (2) only the objective indicators are measured only; (3) the measurement of objective indicators is only the result index. At the same time, the separation of subjective and objective performance is often difficult to carry out in the comparison of the difference of usability. A variety of methods are used to measure the availability of mobile phones. From the point of view of the load, the usability of the mobile phone is measured by the use of the procedural availability index (physiological index). At the same time, the index system of usability is established by combining the subjective index and the different types of objective indexes, and the index system of availability is used by the algorithm. Through mutual verification and prediction, we can solve the problem of effective comparison between different mobile phones by matching index.
In this paper, we determine the dimensions of the index system. (1) the first test of the subjective dimension in the evaluation system has compiled the usability questionnaire of the subjective dimension of the evaluation system. The first test of the project was carried out with 223 subjects. The seven dimensions of the questionnaire were determined by the factor analysis, and the final total of the questionnaires were 40 items, which had good reliability and validity. (2) experiment two verified the effectiveness of EMG as an objective process index. By manipulating the same cell phone, controlling the level of task difficulty and verifying the distinction between electromyography and different load levels, it verifies the effectiveness of EMG as a usability index. Using effective data of 20 subjects, the results are obtained at the level of difficulty of two tasks (simple / sleepy). The average amplitude of the EMG (AEMG), the median frequency (MF), the average power frequency (MPF) on most tasks were significantly different, and the muscle load was larger on the complex task.
Study two based on the BP neural network algorithm, the usability comprehensive evaluation model (3) is built to measure the comprehensive index of mobile phone availability in multiple dimensions. It establishes the basis for the construction of the next model. Using the inter group design, the two groups of subjects use the HTC mobile phone and the Samsung hand machine to complete the operation task respectively, and record the task completion when the subjects are completed. The data of electromyography in the process of between and operation was completed. The results of the test three were compared. On the score of the questionnaire, there was no significant difference between two mobile phones (t=-0569.p=0.572). In the task completion time, two mobile phones had a significant difference in task one, task eight, task twelve and task fourteen. T=-2.420*, p=0.020; t=-3.413**, p=0.001; t=2.538*, p=0.015; t=-3.338**, p=0.002), on AEMG, there is no significant difference between the two mobile phones. For MF, there are significant differences in the other tasks except task 8 and task 14. T=3.526, p=0.001; t=-2.220, p=0.031; t=7.397, p=0.000; t=9.077, p=0.000; t=5.945, p=0.000; t=4.694, p=0.000; p=0.031). Task 7 is more labor-saving than Samsung, and the rest of the tasks are better than Samsung. Significant differences in 2,13,14
(t=4.450, p=0.000; t=6.021, p=0.000; t=2.942, p=0.005; t=6.099, p=0.000; t=3.187, p=0.003; t=8.475, p=0.000; p=0.000; The difference of availability is made in the next step. Based on the BP neural network algorithm, using the nonlinear mapping and using the data as far as possible, the availability evaluation model of the prediction accuracy in 73.3% is established. The weight of the data on the 23 dimensions is obtained, and the availability difference of the two mobile phones is more accurate and objective. The comparison is made.
In this paper, a variety of usability evaluation indexes are synthesized to evaluate the availability of mobile phones, and the limitation of the previous evaluation index is overcome. The evaluation model is established through the algorithm, and the difference in the availability of different mobile phones is compared objectively from the output of the model.
【学位授予单位】：浙江理工大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：B842

【参考文献】