China
Africa
Explore chapters and articles related to this topic
Evaluating Image Fusion Performance
Published in Hassen Fourati, Krzysztof Iniewski, Multisensor Data Fusion, 2016
The Friedman test is a nonparametric equivalent of the repeated-measures analysis of variance (ANOVA) that determines the difference between group means [6]. The Friedman test carries out a multiple comparison test to detect significant differences between two or more algorithms [8]. The algorithms are ranked for each data set. In the case of ties, average ranks are assigned. The Friedman test compares the average ranks of algorithms with the null hypothesis that all the algorithms are equivalent or behave similarly [6].
Machine Learning
Published in Pedro Larrañaga, David Atienza, Javier Diaz-Rozo, Alberto Ogbechie, Carlos Puerto-Santana, Concha Bielza, Industrial Applications of Machine Learning, 2019
Pedro Larrañaga, David Atienza, Javier Diaz-Rozo, Alberto Ogbechie, Carlos Puerto-Santana, Concha Bielza
Friedman test assumptions are that all sample populations are continuous and identical, except possibly for location. The null hypothesis is that all populations have the same location. Typically, the null hypothesis of no difference among the k treatments is written in terms of the medians, ψi $ \psi _i $ . Both hypotheses, H0 $ H_0 $ and H1 $ H_1 $ , can be written as: H0:ψ1=ψ2=...=ψkH1:ψi≠ψjfor at least one pair(i,j) $$ \begin{aligned} \left\{ \begin{array}{l} H_0: \psi _{1} = \psi _{2} = ... = \psi _{k}\\ H_1: \psi _i \ne \psi _j \text{ for} \text{ at} \text{ least} \text{ one} \text{ pair} (i,j)\\ \end{array} \right. \end{aligned} $$
Improving Operator's Conformity with Expectations in a Cognitively Automated Assembly Cell Using Human Heuristics
Published in Stefan Trzcieliński, Waldemar Karwowski, Advances in Ergonomics in Manufacturing, 2012
Marcel Ph. Mayer, Christopher C. Schlick
The subjective rating regarding fatigue was investigated regarding chronological effects of the experiment. For that reason each of the 48 predictions within the order of predictions performed was regarded as factor levels of a factor time. Since not all data is normally distributed, the Friedman-Test on a significance level of a = 0,05 was performed as a nonparametric version of a repeated measures ANOVA.
Exploring the Long-Term User Experience of an Interactive Loading Screen Using UX Curve and QUIS
Published in International Journal of Human–Computer Interaction, 2023
Anping Cheng, Dongming Ma, Hao Qian, Younghwan Pan
The curves of user experience were categorized into deteriorating, stable, and improving based on the difference between the starting and ending point, similar to Kujala, Roto, Väänänen-Vainio-Mattila, Karapanos, et al. (2011). Each curve also was categorized as “negative/neutral/positive start” and “negative/neutral/positive end,” depending if the start and end points were below or above the middle line of the vertical user experience scale. Experience narratives explaining the changes in user experience were content analyzed. Each experience narrative was categorized into negative and positive items. The generated results of QUIS were analyzed by using Friedman Test with SPSS 26.0 software program. After the Friedman Test, Wilcoxon Signed-Rank Tests, a post hoc test, was conducted for the pairwise comparison.
A data-driven iterative refinement approach for estimating clearing functions from simulation models of production systems
Published in International Journal of Production Research, 2019
Karthick Gopalswamy, Reha Uzsoy
In this section, we analyze the numerical results obtained by iterative fitting of functional forms (10) and (13), to quantify the effects of ineffective sampling and model misspecification on the average cost of the production system. The Friedman test (Conover 1980) is a non-parametric test used to compare repeated observations on the same subjects, and uses the ranks of the data rather than their actual values to calculate the test statistic without requiring any distributional assumptions on the data. We perform statistical tests under the null hypothesis that all rankings of the different treatments (in this case, the CF fitting procedures) are equally likely at a significance level of p=0.05. Thus we compute the expected cost obtained using each fitting procedure for each simulation replication of each demand realisation, and observe the relative ranking of the fitting procedures. These rankings then form the input to the Friedman test. The IR algorithm converges in about 8 iterations for the cases with no failures, and within 10 iterations for the cases with failures. To ensure consistent comparisons we fixed the number of iterations at 10 for all experiments.
M-AR: A Visual Representation of Manual Operation Precision in AR Assembly
Published in International Journal of Human–Computer Interaction, 2021
Zhuo Wang, Xiaoliang Bai, Shusheng Zhang, Weiping He, Yang Wang, Dechuan Han, Sili Wei, Bingzhao Wei, Chengkun Chen
The results of Friedman test show that there are significant differences among all items (Q1: χ2(2) = 22.917, p =.008; Q2: χ2(2) = 21.147, p =.009; Q3: χ2(2) = 20.113, p =.004; Q4: χ2(2) = 19.001, p <.001; Q5: χ2(2) =18.014, p =.002; Q6: χ2(2) =21.008, p <.001; Q7: χ2(2) =14.667, p <.001; Q8: χ2(2) =11.461, p <.001). The above analysis shows that the three affect the evaluation of participants from eight aspects (i.e., enjoyment, focus, self-confidence, natural intuition, feasibility, effectiveness, availability, and comprehensibility), and affect the cognitive efficacy of participants on visual information.