Search

The Exstatix statistical analysis software package by K. C. Killion for use with Macintosh computers is evaluated. In evaluating the package, the framework developed by C. J. Ansorge et al. (1986) was used. This framework encompasses features such as transportability of files, compatibility of files with other Macintosh software, and ability to merge results with word processing and desk top publishing applications. The review is from the perspective of the educational researcher who wishes to use the package for uncomplicated, quick-and-dirty analysis of relatively small univariate data sets and export the output to reports. Specific features covered in the evaluation include documentation, ease of use, running programs, program output, accuracy, and other features. Strengths of Exstatix are discussed in the areas of statistical decision making, assumption testing, and regression analysis. Weaknesses of the packages include problems with importing and exporting functions, report generation, analysis of variance applications, and add-ons. Sample output of tables and graphs are provided. (TJH)

Electrical discharge machining (EDM) is a process for shaping hard metals and forming deep complex shaped holes by arc erosion in all kinds of electro-conductive materials. The objective of this paper is to study the influence of operating parameters of tungsten carbide on the machining characteristics such as surface quality, material removal rate and electrode wear. It is followed by optimizing the machining condition for confirmation test purposes. Then, the conformation test needed to evaluate the error margin between predicted result by software and confirmation result by experiment in terms of the machining characteristics. The effectiveness of EDM process with tungsten carbide, WC-Co is evaluated in terms of the material removal rate, the relative wear ratio and the surface finish quality of the workpiece produced. It is observed that copper tungsten is most suitable for use as the tool electrode in EDM of WC-Co. Better machining performance is obtained generally with the electrode as the cathode and the workpiece as an anode. In this paper, a study was carried out on the influence of the parameters such peak currrent, power supply voltage, pulse on time and pulse off time. The surface quality that was investigated in this experiment was surface roughness using perthometer machine. Material removal rate (MRR) and electrode wear (EW) in this experiment was calculated by using mathematical method. The result of the experiment then was collected and analyzed using STATISTICA software. This was done by using the design of experiments (DOE) technique and ANOVA analysis. [ABSTRACT FROM AUTHOR]

Injection moulding parameters that simultaneously satisfy green part qualities (green defect, green strength and green density) have been optimised using L27 (313) Taguchi orthogonal array. Moulding variables involved in the optimisation are the moulding pressure, moulding temperature, mould temperature, holding pressure, moulding rate, holding time and cooling time. The analysis of variance (ANOVA) is employed to determine the significant levels (α) and contributions of the variables to the green qualities. Results show that the mould temperature and holding time is highly significant simultaneously to the green qualities, while the holding pressure and cooling time does not show any significance. Besides that, densification of the compact moulded with the optimised moulding parameter is presented in this paper. With sintering temperature ranging from 1340 to 1400 °C while heating rate and dwell time remains at 10 °C/min and 4 hours respectively demonstrates that the optimised moulding parameter is enabled to produce a sintered part with a density which is very close to the solid density of SS316L. [ABSTRACT FROM AUTHOR]

Manufacturing enterprises consist of people, things, and information. Peoples are hired for performing various functions such as; marketing, design, purchasing, inventory control, inspection, machining, management, safety, service, and security. Things range from factories, to machines, materials, transporters, computers, warehouses, vendors of components, and utilities. Information is according to marketing requirements, product design, and Production systems and operations, Production processes, human resources, supplier chain systems, and general management. All these elements constitute part of the manufacturing enterprise. A Manufacturing organization as a system can have functional sub-systems, such as a production sub-system, a marketing sub-system, management sub-system, management and information sub system, and labor sub-system. The scope of this paper is limited to design and development of a model for Competitive Production Management System (CPMS). A Delphi approach along with 'ANOVA' has been used to separate out significant and non-significant factors affecting in the models, resulting in a comprehensive, compact sub-system design. Finally, a model designed and developed for CPMS has presented at the end. An Axiomatic Design approach selected for design and development of CPMS. At the end, comprehensive Mathematical model for CPMS presented. [ABSTRACT FROM AUTHOR]

This paper considers robust mean-variance portfolio selection problems including uncertainty sets and fuzzy factors. Since these problems are not well-defined problems due to fuzzy factors, it is hard to solve them directly. Therefore, introducing chance constraints, fuzzy goals and possibility measures, the proposed models are transformed into the deterministic equivalent problems. Furthermore, since it is difficult to solve them analytically and efficiently due to nonlinear programming problems, the solution method is constructed introducing a parameter and doing the equivalent transformations. [ABSTRACT FROM AUTHOR]

The objective of the paper is to obtain an optimal setting of turning process parameters--cutting speed, feed and depth of cut, which may result in optimizing tool life of TiC coated carbide inserts while turning En24 steel (0.4 % C ). The effects of the selected process parameters on the tool life and the subsequent optimal settings of the parameters have been accomplished using Taguchi's r design of experiments approach. The results indicate that the selected process parameters significantly affect the mean and variance of the tool life of the carbide inserts. The percent contributions of parameters as quantified in the S/N pooled ANOVA envisage that the relative power of feed (8.78 %) in controlling variation and mean tool life is significantly smaller than that of the cutting speed (34.89 %) and depth of cut (25.80 %). The predicted optimum tool life is 20.19 min. The results have been validated by the confirmation experiments. [ABSTRACT FROM AUTHOR]

Identification of nonlinear dynamic black box models involves structure detection of nonlinear system (i.e. selecting the regressors that have the most contribution to the output and the regressor function) and finally estimation of model parameters. As the NARX representation can describe many nonlinear dynamic models, it will be used here as the desired structure. It should be noted that when the order of the system increases, even for moderately complex systems the number of candidate terms becomes very large. So, structure detection is necessary in order to have an efficient description of the dynamic systems. In this paper, a new method for selecting regressors with the most contribution to the output and finding an efficient representation of nonlinear dynamic systems is presented. The purposed method, named ANOVA-BSD, is based on the combination of analysis of variance and suboptimal bootstrap algorithm. The anticipated structure takes the advantage of nonlinear ARX polynomial to model different nonlinearities of the system, such as sine and cosine functions. The proposed method is tested on two different systems and simulation results show that ANOVA-BSD effectively reduces model complexity without any noticeable loss in the accuracy. [ABSTRACT FROM AUTHOR]

There are two main reasons why zeros are of particular interest in count data models. First, empirically, their fraction is often too high to be compatible with a standard underlying count data model (we also speak of excess zeros then). Second, theoretically, zeros often reflect corner solution outcomes in economic choice models. In such cases, the process generating zeros might depend on other driving forces than the process for strictly positive outcomes, making it informative and relevant to distinguish between elasticities at the intensive and extensive margins. This issue is reminiscent of the debate in the limited dependent variable literature on the appropriateness of the Tobit model as opposed to so-called two-part models (Cragg, 1971, Duan et al., 1993). [ABSTRACT FROM AUTHOR]

This book discusses specification and estimation of regression models for nonnegative integers, or counts, i.e., dependent variables that take the values y = 0, 1, 2. . . without explicit upper limit. Regression analysis, narrowly defined, attempts to explain variation in the conditional mean of y with the help of variation in explanatory variables x. If the mean function is embedded in a probability distribution, one obtains a full conditional probability model of y given x. [ABSTRACT FROM AUTHOR]

This chapter contains two case studies. In the first, two rubber machines were malfunctioning and causing a production bottleneck. Standard screening using fractional factorial methods were applied to identify the cause and suggest a prompt and successful remedy. One of the associated factors used in the study was not a setting on a machine but rather a way of stating policy to employees. In the second study, an innovative design of experiments methods called low cost response surface methods (LCRSM) was applied to develop a surface prediction of strength and insertion effort for snap tabs. Formal optimization of the resulting surface models permitted the doubling of the strength with small increase in size. [ABSTRACT FROM AUTHOR]

This chapter describes the application of so-called response surface methods (RSM). These methods generally result in a relatively accurate prediction of all response variable averages related to quantities measured during experimentation. An important reason why the predictions are relatively accurate is that so-called “interactions” which relate to the combined effects of factors are included explicity in the predicted models. Three types of methods were presented. Box Behnken designs (BBDs) were argued to generate relatively accurate predictions because they minimize so-called “bias” errors under certain reasonable assumptions. Central composite designs (CCDs) were presented and explained to offer the advantage that they permit certain level adjustments and can be used in two-step sequential response surface methods. In these methods, there is a chance that the experimental will stop with relatively few runs and decide his or her prediction model is satisfactory. The third class of experimental designs presented is the expected integrated mean squared error (EIMSE) designs which are available for a variety of numbers of runs and offer predictive advantages of Box Benken designs. The EIMSE criteria is also used at the end to clarify the relative prediction errors and to help method users decide whether a given experimental design is appropriate for their own prediction accuracy goals. [ABSTRACT FROM AUTHOR]

This chapter discusses the use of nonparametric productivity analysis in measuring and comparing the relative performance of production units. The basic concepts of productivity analysis and two popular methods, Data Envelopment Analysis (DEA) and Free Disposal Hull (FDH) model are reviewed. An example from the empirical study of Scheel (1999) is considered in order to illustrate the practical application of DEA. This concrete data analysis is about the efficiency of 63 agencies of a German insurance company. In order to illustrate how FDH works, the Manufacturing Industry Productivity Database from the U.S. National Bureau of Economic Research is discussed.

This chapter presents a thorough presentation of the computational aspects related to stable laws. Stable laws were introduced by Levy (1925) during his investigations of the behavior sums of independent variables. Stable distributions often give a very good fit to empirical data. They are valuable models for data sets covering extreme events, like market crashes or natural catastrophes. They are a useful tool in the hands of an analyst working in finance or insurance.