Jing Zhao

The Comparison of Three Robust Statistical Methods in Proficiency Testing

Three commonly used robust statistical methods are introduced in this paper, namely quartile method, algorithm A, and Q/Hampel method. Two application examples are presented, and results of the three methods are compared. For data set of symmetrical distribution and small proportion of outliers, quartile and algorithm A are applicable. For non-symmetrical shape data set or a large proportion of outliers (>20%), Q/Hampel method is recommended. Code for Q/Hampel method is also provided to help quick calculation for this method.

Parameter Estimation of Exponential Distribution

In this paper a new parameter estimation method is given in view of another common situation of exponential distribution (stipulate monitor time, in each monitor time observing the number of expired sample, namely do not observe exact time). The properties of this estimate is also discussed. Simulation is given to illustrate the feasibility of this method. Keywords—multinomial distribution; exponential distribution; dichotomy

A Regression Equation Model for Height and Weight Prediction

Height and weight prediction has been a popular problem in ergonomics study. In this paper, we reduce the dimension by principal component analysis and choose the best regression equation using various statistical criterion such as Residual Mean Square (RMSq), Mallow Cp and Akaike information criterion (AIC). Finally, compared with the real value, we analyze the fitting accuracy of the regression equation we proposed.

A Study on Young Men’s Upper Body Characteristic Indexes in China Considering Human Ergonomics

This paper takes analysis of the upper body type of young men in China aiming to give some implications for shirt design. According to the characters of sample shape, extract the key control variable to conduct size classification, and at the same time, take a detailed analysis to several important area of the upper body shape. The specific content is as follows: (1) Twenty six upper body measurements obtained from body scanner were designed for analysis. Firstly, R-type cluster analysis was carried out on the measured data, and it is concluded that the twenty six indexes could be clustered into five classes. (2) Then a multiple linear regression model was established which could be applied to the predict the specific body parts. In order to get the optimal predict results, next we choose the best regression equation using various statistical criteria such as Residual Mean Square (RMSq), Mallows Cp. From the best regression model, detail size of body can be deduced. The data from the prediction model are precise, which can make the garment pattern more precise and suitable. The results can also provide the body foundation for some empirical formulas in shirt pattern design.

Research on the Characteristic and Gender Differences in Adult Hand Shape in China

The purpose of this study was to develop some new knowledge for structuring a sizing system for men’s and women’s hand tools, gloves and so on. Firstly, the paper per-processed the obtained data and then conducted necessary descriptive statistical analysis to get a general knowledge. It described their hand shape characteristics and how they differ among different groups classified according to the age bracket. The results show that with age growing, the length of hand, the breadth of hand and index finger breadth, proximal for men are first increasing then decreasing, and the most obvious representative are 20-year-old and 50-year-old boundaries. With age growing for women, there are little changes for most indexes of hand measurements. Additionally, we conduct univariate t-tests and multivariate discriminant analyses for some age bracket to analyze gender differences in hand shape in a large sample of individuals. The results showed that overall men had longer and broader hands than women for a given stature. After normalization of the measurements by height, men and women were found to differ significantly in two hand shape variables. Classification by gender using absolute values was correct at least 92% of the time. After normalization of the measurements by hand length, classification was correct at least 88% of the time. This study demonstrated that female hands were not simply scaled-down versions of male hands but rather differ in a number of shape characteristics.

Research on the Characteristics of Body Height and Weight in Eight Countries

The data about body height and weight is the basic data information in anthropometric study. Firstly, this paper gave the direct comparison among these countries based on analysis of variance (ANOVA). The results of ANOVA showed that gender and country were both significant factors for analyzing body height or weight in these countries. No interaction effect exists between the two factors. Based on the average data and percentile values of weight we can conclude that these eight countries fall into three groups: The Netherlands and United States had significantly larger weight than the other six countries; China and India had the smallest weight. As expected, men are taller than women. In general, the rank order of samples with regard to height was similar in men and women. Men and women in The Netherlands were the tallest and people in India the shortest. There was a strong linear association between height and weight in men and women. Secondly, height and weight inference were conducted based on the ergonomic data in these countries. The relation between human height with the main body indexes was established using classical regression analysis. The obtained results could be a reference for designing the export clothes and food to these countries.

Visual Analysis on Macro Quality Data

Quality inspection and quarantine processing is an important process with extensive business impacts. Data visualization methods failed to fully combine standards and certifications or accreditation. While inspection, quarantine and other quality related measures are hard to analyze due to the multi-factor dimension and therefore it is difficult to reflect the macro quality status. Thus, in order to integrate quality data from quality supervisory inspection and quarantine departments, this paper propose three visualization methods to analyze quality data, such as tree models and histogram visualization, map and histogram visualization, the models implement quality visualization system based on the methods above to realize the comprehensive analysis of macro quality data.

Decision Rules Mining with Rough Set

Decision rule has been widely used for its briefness, effectiveness and favorite understandability. Many methods aiming at mining decision rules have been developed. Rough set theory. Unfortunately, data is split between multiple parties in many cases. And privacy concerns may prevent these parties from directly sharing the data. This paper addresses the problem of how to securely mine decision rules over horizontally partitioned data with rough set approach. This paper integrates a general framework rather than a very specific solution for mining decision rules with rough set approach when privacy is concerned and data is horizontally partitioned.

Study on the Relationship Between BMI and Body Shape Characteristics Indicators

Body Mass Index is widely used international to measure the extent of the body fat or thin, and it’s also looked as healthy standard. This paper studies the relationship between Body Mass Index (BMI) and body shape characteristics through the anthropometric data of statistical analysis. The density function figures of the three BMI types corresponding to each indicator are given to discover the relationship between the BMI classification and body shape characteristics. Then Chest circumference, waist circumference, hip circumference and other six indicators are selected as body shape characteristics indicators, which has significant relationship with BMI. Moreover, the multiple regression model is built between them, and by using the model to forecast the BMI value, the accuracy of the model is verified.

Foot Anthropometric Data for Young Chinese Population

As the rapid development of whole society of China, consumers pay more attention to the comfort and fit degree of garment accessory like footwear. Precise foot size can help good design for footwear and other related accessory. In this paper, we investigated the foot anthropometric data of 664 Chinese young women and 596 Chinese young men, with ages from 18 to 25 years old, by using 3D scanning equipment. Descriptive statistics of each data set was obtained. Then using ANOVA analysis the foot length and foot breadth among different ages are compared. The three ANOVA assumptions are verified for each data set. Only the data of foot breadth of women has inequality variance, so Welch test is used for this data. The final results show that women foot data is not different among ages, but men foot data is significantly different among ages.