China
Africa
Explore chapters and articles related to this topic
The Data Analysis and Result Interpretation Correlating Human Error and Technology Advancement in Nuclear Operation
Published in Jonathan K. Corrado, Technology, Human Performance, and Nuclear Facilities, 2023
I used the t-test to determine whether a significant difference between the two groups (incidents determined to be caused by human error as a result of technological advances and those determined not to be caused by human error as a result of technological advances) existed. An independent samples t-test determines whether two independent groups are different from each other, and the two groups are not related, so an independent samples t-test can be used in this case. Before conducting the t-test, since it is a parametric test and requires normally distributed data, the outcome variable (costs of incidents) was checked for normality and homoscedasticity of variance by examining the distribution of the data using histograms and Q-Q plots [1].
Biostatistics and Bioaerosols
Published in Harriet A. Burge, Bioaerosols, 2020
Lynn Eudey, H. Jenny Su, Harriet A. Burge
Example 4. Comparing the locations of two independent populations, small samples. If it is not possible to collect large samples from each of the two populations, a t-test can be used to compare the population means. The assumptions of the t-test are that independent samples are taken from two normally distributed populations and that the two populations have the same variance. The form of the test statistic (Step 3) is the same as it is for the test above (except that it is called T instead of Z) and we pool the estimates of the standard deviation. The T statistic is compared with a t-distribution with degrees of freedom = (n1 + n2 − 2). If both of the populations are normally distributed but the variances are not equal then theoretical problems arise. There are approximate methods for this situation which use a T statistic (with a different denominator) but refer to a Student’s t-distribution with an adjusted degrees of freedom.
Ideation
Published in Walter R. Paczkowski, Deep Data Analytics for New Product Development, 2020
Another level of analysis could be done, data permitting. For each attribute, the difference between the mean performance of your product and the mean of each of the competitors’ mean performance can be determined for each attribute. The difference should, of course, be statistically tested. The recommended test is Dunnett’s multiple comparison test. In multiple comparison tests in general, there is an issue of performing more than one statistical test on the same data. The standard level of significance, which is the probability of falsely rejecting the Null Hypothesis, used in, say, a t-test is α = 0.05.18 This is sufficient when one test is performed to determine a difference. When more than one test is performed, however, it can be shown that the probability of falsely rejecting the Null Hypothesis, H0, is greater than 0.05. In fact, Pr(Falsely Rejecting H0) = 1 − (1 − α)k where k is the number of tests. Table 2.3 shows what happens to this probability for different values of k.
Integrated mixture model and ensemble learning geographic object-based image analysis for road network extraction
Published in Journal of Spatial Science, 2023
Elaveni Palanivel, Shirley Selvan
The features are then analysed statistically to determine the significant features. This helps reduce the complexity of the proposed algorithm by reducing the amount of data that needs to be processed. Also, this contributes to improved accuracy of the output. The p-value of the features is calculated and only the most significant of them are chosen as inputs for further classification. Initially, a population of 1000 patches is compiled. All the listed features are extracted from the patches to create a database. Both T and Z tests are implemented. The T-tail test is calculated for a small sample size and the Z-test is for a larger sample size. The p-value of the features is listed in Table 2. 25 samples are used to carry out the T-test and all 1000 samples are used to carry out the Z-test. When the p-value is less than 0.05, the null hypothesis is rejected in T-test. In the case of the z two-tailed test, the z value has to be greater than 1.96 or less than 1.96 for the null hypothesis to be rejected.
A study on securing model usefulness through geographical scalability testing of winter weather model developed with big traffic data
Published in Transportation Planning and Technology, 2022
In the previous section, a portion of the preliminary data analyses was carried out only for two vehicle classes due to the similarity of total traffic with passenger cars. However, this section presents all estimated model parameters for all three vehicle classes, including total traffic. Table 2 shows parameter estimates for the modelling site using Eq. (1) and statistical test results. The squared multiple correlation coefficient values for all models are more significant than 0.99. In other words, the estimated models are well-fitted with the data of the modelling site. The value of the test assesses the overall adequacy of the model. This statistic says the model structure is statistically valid at a better than 0.001 significance level (Roh and Sharma 2018; Roh 2021a). t-statistic shows the statistical significance of individual coefficients. The significance levels of t-tests are delivered using a significance system ap<0.001, bp<0.01, cp<0.05. The exact t value for each coefficient is included in the second column from the right in Table 2. Standard error and p-value for each parameter estimated are included in the third and first columns, respectively, from the right in Table 2.
Exploring the impact of project size on design-bid-build and design-build project delivery performance in highways
Published in Construction Management and Economics, 2021
Phuong H. D. Nguyen, Dai Q. Tran, Sai P. K. Bypaneni
The use of t-test depends strictly on the satisfaction of the required assumptions: (1) normal distribution of the populations and (2) equality of variances between two groups. If the D-B-B and D-B samples are normally distributed and have equal variances, the two-sample t-test is selected. If the first assumption is violated (i.e. the samples are not normally distributed), the Mann–Whitney U-test, a non-parametric test equivalent to t-test, is utilized to examine the difference between mean values of two samples. If the second assumption is violated (i.e. the samples have unequal variances), a variation of t-test, the Welch’s t-test, is utilized to compare mean values between two samples. Prior to conducting the hypothesis testing process, a confidence level of 95% (α = 0.05) was set to calculate the significant value (p-value) for the subsequent testing processes. The null hypothesis for the used tests was that the mean values of performance metrics (i.e. cost growth, schedule growth, and construction intensity) between D-B-B and D-B projects are equal within the five groups of project size. If the p-value is less than 0.05, the difference in mean values between D-B-B and D-B projects is statistically significant.