N.A

  The purpose of this article is to review method(s) of constructing orthogonal arrays (both symmetric and mixed) by exploiting the concept of resolvable (symmetric) orthogonal arrays and resolvable mixed orthogonal arrays. Several series of symmetric and mixed orthogonal arrays obtained in the literature by using the notion of Kronecker product and Kronecker sum of OAs have also been described. Some general methods of obtaining orthogonal arrays and mixed orthogonal arrays have also been given. Applications of orthogonal arrays, mixed orthogonal arrays and resolvable orthogonal arrays have been described. Key words: Symmetric and Mixed Orthogonal array; Resolvable orthogonal array; Nested Orthogonal Array; Supersaturated design; Irregular fractions; Balanced repeated replications.

  For reliable policy planning at micro-level, estimates of crop yield at small area level, say block level, is required. Application of existing crop-cutting methodology would not be feasible in view of prohibitive cost involved. One possible alternative is to employ ?Fuzzy regression methodology?. Accordingly, in this paper Possibility and Necessity measures for obtaining reliable fuzzy estimates of crop yield have been thoroughly studied. Estimation of parameters is carried out using ?Fuzzy least-squares? procedure. As an illustration, the methodology is applied to Pearl Millet crop yield data in order to build block level estimates for Bhiwani district, Haryana based on farmers? estimates at the same level. Performance evaluation criterion is used to compare results of Possibility and Necessity approaches at optimal value of fitness level. Key words : Crop-cutting experiments, Possibility measures, Necessity measures, Fuzzy least-squares, LINGO, Pearl Millet.

  The study ?Technical Efficiency of Dairy Farms in Tamil Nadu? was carried out to evaluate dairy farm households in terms of efficiency of milk production using stochastic frontier production methods. The data for the study comprised of fixed investments on dairy farms, quantity and price of feeds and fodders fed to individual animals, labour utilization pattern, veterinary and miscellaneous expenses, quantity of milk produced and price realized, etc. collected from 160 sample households across flush and lean seasons for the year 2002-03. The coefficients for the value of green fodder and concentrate were found to be statistically significant with a relatively higher magnitude implying their greater and significant role in crossbred cow milk production. The technical efficiency of crossbred cow farms ranged from 72.30 to 97.90 per cent with an average of 82.10 per cent. The study indicated that there existed a scope to increase milk production of an average farm by 16.32 per cent for crossbred cows and 14.04 per cent for buffaloes without incurring any extra expenditure on these farms. Key words: Technical efficiency, Stochastic frontier production function.

  In regression models when the errors are correlated, the sample residuals contain some information about the future observations. This information, which is generally ignored, has been used in this paper to improve the precision of predicting the post-sample observations. The best linear unbiased predictor for an m-equation linear SURE model has been obtained under the assumption that the errors in each equation follow first-order autoregressive scheme. The gain in efficiency of the proposed predictor over the usual generalized least squares predictor has been obtained and the method is illustrated for a two-equation acreage response model. Small sample properties of the predictor have been studied by using a Monte-Carlo experiment. Key words: Seemingly unrelated regression equations, Best linear unbiased predictor, GLS predictor, Autoregressive errors.

  The multidimensional modeling tools, which are of recent origin in the field of data warehousing, have vast potential for development of Web-enabled decision support system through deployment of Web based multidimensional cubes. In this article, an attempt has been made to present experiences of the authors related to different processes and techniques involved in design and development of multidimensional cubes in reference to agricultural sector during implementation of Central Data Warehouse (CDW) at Indian Agricultural Statistics Research Institute (IASRI), New Delhi. The features of multidimensional model from the user?s perspective have been highlighted to demonstrate the power of this On-line Analytical Processing (OLAP) technology. Key words: Analysis, Data warehouse, Modeling, Multidimensional cubes, OLAP.

  N.A

  N.A

  N.A

  N.A

  Hansen and Hurwitz (1946) suggested a technique for eliciting responses from a subsample of the non-respondents. Khare (1987) applies this procedure of subsampling in stratified sampling and discussed the problem of optimum allocation in presence of non-response. When more than one characteristics are under study, it is not possible to use the individual optimum allocations for one reason or the other. In such situations, some criterion is needed to work out an acceptable sampling fraction which is optimum for all characteristics in some sense. In this paper, the problem of determining the optimum allocation and the optimum size of subsamples to various strata in multivariate stratified sampling in presence of non-response is formulated as a Nonlinear Programming Problem (NLPP). A solution to this problem is obtained using Lagrange multipliers technique. Explicit formulae are obtained for the optimum allocation and the optimum sizes of the subsamples. Key words: Non-response, Multivariate stratified sampling, Optimum allocation, Optimum size of subsamples, Nonlinear Programming Problem.

  Agricultural Statistician Network provides dynamic working linkages among the statisticians with emphasis on research information exchange, resource sharing and optimizing response time for addressing methodology related problems and foster fellow feelings among the group with cost effective communication media. The website is accessible across the world at the address www.iasri.res.in/asn/ and provides an opportunity to enhance the image of organizations and the statisticians at the international platform. It lays emphasis on providing a unified base on various professional activities of the agricultural statisticians. Individual?s research expertise is exposed to the international community thus providing opportunity for global collaboration. System has features like chat, discussion group, notice board, search, online registration and online data management for effective communication. Web site has been designed and developed using standard three-tier architecture. Site has been developed using Active Server Pages (ASP) for server side programming, Hyper Text Markup Language (HTML) and VBScript for client side interface and validation and Microsoft Access for Database management. Key words: Agricultural Statistician Network, Agricultural Statistics, ASN, Online Information System.

  Statistical Package for Agricultural Research (SPAR 2.0) has been developed under the Windows platform for handling statistical analysis of agricultural experimental research data. The package has eight modules to carry out different aspects of analysis such as diallel analysis, path analysis, discriminant analysis, cluster analysis, line × tester analysis, stability analysis etc. useful for plant breeding and Genetics research studies. This package is also useful for teaching the subject of Genetical Statistics to the Post Graduate students and researchers in statistics with special interest in Plant and Animal Sciences. There are no high end requirements for the successful execution of the package, a normal system with minimum configuration is enough to run this package.

  Classification is an important research topic in the field of data mining and knowledge discovery. There have been many data classification methods including decision tree methods, statistical methods, neural networks, rough sets, etc. A reduct in the rough set theory refers to a set of dominant attributes in the dataset. A dataset may have zero, one or multiple reducts. A classification problem utilizing information contained in a single reduct is well examined in rough set literature. However, it means ignoring the available knowledge from the multiple reducts. An approximate core is proposed as an important tool to deal with the datasets which are having multiple reducts. In this paper, Forest cover type, a large benchmarking dataset having multiple reducts is used for experiments. The performance parameters - accuracy, complexity, number of rules and number of attributes in the resulting classifiers are compared among various algorithms employed. The results using approximate core are comparable with the other published results for this dataset. Key words: Rough sets, Decision tree, RDT model, Approximate core, Forest cover type.

  N.A