The importance of spatio-temporal modeling is increasing day by day in many areas with the growing accessibility of spatio-temporal data. The spatio-temporal time series data exhibits spatio-temporal relationships among themselves. Therefore, it is essential to recognize those relations and incorporate it to the models for getting better forecast. On the other hand, forecasting based on fuzzy techniques are very much useful for the imprecise data like rainfall, temperature etc. In the last three decades, various univariate methods have been developed for forecasting the time series based on fuzzy techniques. Since these methods are only valid for the single time series, it cannot utilize for spatio-temporal time series. To overcome this limitation, a new forecasting method has been developed using fuzzy techniques in conjunction with the k-means clustering. The proposed method has been empirically illustrated for forecasting the annual precipitation data of six districts of West Bengal. The results from this study, reveals the Keywords: Fuzzy technique, K-means clustering, Rainfall, Spatio-temporal time series.

  This article deals with generalized row-column designs when there are two sets of treatments, one set consisting of test treatments and the other of control treatments called Bipartite Generalized Row-Column Designs. The two sets are disjoint in the sense that there are no treatments in common between the two. The interest here is to estimate the contrasts pertaining to test treatments vs. control treatments with as high precision as possible. designs ensure that all the contrasts pertaining to test vs. control are estimated with less variance in comparison to those pertaining to test vs. test treatments. Keywords: Row-column design, Disjoint set, Test treatments, Control treatments, Bipartite.

  Paul and Bhar (2011) advocated the use of M-estimation methods to address the issue of dealing with outliers in designed experimental data. An attempt has been made to elucidate the efficacy of this method over the regular ANOVA method using real time horticulture perennial crop experimental data. Results fortified the efficacy of robust methods while dealing with outliers as revealed by the three to five fold reduction in error sum of squares coupled with acceptable probability values for all the characters. Thus this study calls for adoption of robust ANOVA approach while dealing with outliers in perennial horticulture crop experiments in future research. Keywords: Outlier, Papaya, Robust ANOVA.

  Latin hypercube designs (LHDs) are commonly used in designing computer experiments. In recent years, several methods of constructing orthogonal Latin hypercube designs have been proposed in the literature. In this article, two new series of second order orthogonal Latin hypercube designs for six factors have been given. Keywords: Latin hypercube designs, Orthogonal, Second order, Computer experiments.

  The purpose of this study is to identify the most important factors affecting poverty in rural Odisha by using household level 68th round primary data (2011-12) collected by the National Sample Survey Office (NSSO) on consumer expenditure. In this study a logistic regression analysis is undertaken where the dependent variable is a dichotomous variable (Y), coded as Y = 1 for the household below official poverty line and Y = 0 for the household above poverty line and thirteen explanatory variables are age and sex of the head of the household, education, marital status of the the head of house hold, family size, at least one salary earner in the household, social group of the household, total land possessed by the household, total cultivated land of the household, number of dwelling units in the household, source of energy for lighting and cooking and percentage of monthly cultivated land of the household, number of dwelling units in the household, source of energy for lighting and cooking and percentage of monthly cultivated land of the household, number of dwelling units in the household, source of energy for lighting and cooking and percentage of monthly cultivated land of the household, number of dwelling units in the household, source of energy for lighting and cooking and percentage of monthly per capita expenditure under food items. It was found that age, marital status and general education level of the head of the household, family size, at least one salary earner in the household, social group of the household, source of energy for lighting and percentage of monthly per capita expenditure least one salary earner in the household, social group of the household, source of energy for lighting and percentage of monthly per capita expenditure under food items are significant predictors affecting poverty in rural Odisha. Keywords: Poverty, Logistic regression, National Sample Survey Office, Odisha. Keywords: Poverty, Logistic regression, National Sample Survey Office, Odisha. Keywords: Poverty, Logistic regression, National Sample Survey Office, Odisha. Keywords: Poverty, Logistic regression, National Sample Survey Office, Odisha.

  The demand for acceptable disaggregated level statistics from sample surveys has grown substantially over the past decades due to their extensive and varied use in public and private sectors. Basically, it is the main endeavor of ?Small area estimation (SAE)? approach to produce sound prediction of a target statistic for small domains to answer the problem of small sample sizes. The traditional survey estimation approaches are not suitable enough for generating disaggregate or small domain level estimates because of sample size problem. The SAE techniques therefore provide a feasible way to produce the reliable estimates at disaggregate level from the existing survey data.This paper explores a spatial dependent aggregated level Hierarchical Bayes (HB) model for SAE to estimate the yield for paddy (green) crop at district level in the state of Uttar Pradesh in India. The approach uses survey data from the Improvement of crop statistics (ICS) scheme collected by National Sample Survey Office (NSSO) and linked with Population Census. A considerable gain has been obtained while exploiting spatial information in aggregated level small area model. Keywords: Aggregated level; Small area estimation; Spatial Information.

  A method of identifying subset of outliers in presence of masking has been developed for incomplete Multi-Response design. Design is composed of two sets of experimental units. Different numbers of response variables are observed from these two sets. A Conditional Cook?s Statistics in block design for incomplete multiresponse experiments has been developed for identification of outliers in presence of masking. The developed statistic has been illustrated with a real life data set. It has been shown that outliers in presence of masking can distort the overall conclusion from an experiment. Keywords: Incomplete Multi-response experiments; Outlier; Masking; Cook-statistic.

  The lactic acid bacteria, Lactococcuslactis, is used as a starter culture(s) in food fermentation(s). The specific growth information can be predicted for such starters, that can be practically valuable in exploiting lactic bacteriafor required fermentations. These bacteria are responsible to produce various metabolites. The metabolites of interest are produced from these starters under a set of known growth conditions. Their growth can be modeled using selected mathematical functions. These functions can be used in determining the parameters like specific growth rate and lag time of the organism under defined environmental conditions. In this study, out of the three functions (Gompertz, Logistic and Richards functions) used, Gompertz function was selected. The model gave out constants that were used to obtain biological parameters for its growth. The main aim was thus to ascertain a particular function and the Goodness of fit (from R2 values) that was measured from the fitted growth data. Durbin-Watson test was used to test the residuals dependency by autocorrelation. A larger number of fd values was the criteria to suggest if either the data plotted on Logistic or Gompertz function (3 parameters functions), was more helpful. The Gompertz function was found to be superior to Logistic function (both being three parameter function), against Richards function (a four parameter function) that had inherent and variable degree of skewness for testing the Gompertz and Logistic functions. The selected function, Gompertz function, was then derivatized and biological parameters were calculated, from the constant values so obtained. Thus the kinetic data with respect to time was resolved into an easy way to calculate biological parameters in terms of the simple equations. This methodology can be extended to lactic acid bacteria producing various other metabolites of interest. Keywords: Microbial modelling, Fermentation, Non-linear growth model, Lantum points.

  Classification is the data mining task of assigning the objects to one of the several predefined categories. It is a predictive modelling task in which a model is built for the target variable as a function of the explanatory variables. It is also called the supervised learning since the training dataset has records with predefined labelled classes. These labelled training records supervise the learning of the classification model. The various class labels can be represented by discrete values where the ordering among the values has no meaning. There are many well established techniques for classification, out of which decision tree technique is a very important and popular technique from the machine learning domain. C4.5 is a well-known decision tree algorithm used for classifying datasets which is available in all data mining software. Since it is an important algorithm for inducing the decision trees and generating the rules precisely from the datasets, it is highly used by the data mining and machine learning community. To provide an online platform to the users for applying the algorithm on their datasets without installing any data mining software, a web based software for rule generation and decision tree induction using C4.5 algorithm is developed. The visualization in the form of tree structure enhances the understanding of the generated rules. Keywords: Classification, Data mining, Predictive modelling, Decision tree, Visualization.

   N.A

  The research work presented in this paper is motivated by a real life scenario in the context of agricultural experiments. It is believed that the The research work presented in this paper is motivated by a real life scenario in the context of agricultural experiments. It is believed that the The research work presented in this paper is motivated by a real life scenario in the context of agricultural experiments. It is believed that the neighboring plots in a Block Design or in a Latin Square Design tend to influence each other in terms of the mean yield through the 'effects of the treatments' applied in these plots. We contemplate a linear model and study its analysis in considerable details. Keywords: Block designs, Latin square designs, Direct treatment effects, Neighbor treatment effects, Left neighbors, Right neighbors, Diagonal neighbors, Linear model, ANOVA.

  Auxiliary information is often used to improve the precision of estimators of finite population total. Calibration approach (Deville and Sarndal, 1992) is widely used for making efficient use of auxiliary information in survey estimation. Aditya et al. (2016) proposed regression type estimators of the population total using the calibration approach under the assumption that the population level auxiliary information is available at secondary stage unit level under two stage sampling design. In this paper we have proposed an improved variance estimator of the regression type estimator proposed by Aditya et al. (2016) using higher order calibration approach (Singh et al., 1998). We carried out limited simulation studies to demonstrate the empirical performance of proposed estimators. Our empirical results show that the proposed estimator performs better than the usual estimator of variances of the regression type estimator (Aditya et al., 2016). Keywords: Auxiliary information, Calibration approach, Regression type estimator, Higher order calibration, Two stage sampling.

  The morphological properties of agricultural commodities, such as size, shape, colour etc. are required in various agricultural applications including quality assessment, crop/ variety identification and design of machinery. Manual measurement of these attributes is arduous, time taking and often not so accurate. An approach based on image processing has been developed as computer software. This approach is a rapid, non-invasive, and quantitative. The software reads an image, taken using flatbed scanner, preferably in TIFF (tagged image file format), and instantaneously provides the physical parameters to an accuracy of 0.1 mm. The aim of the present work was to ascertain accuracy and precision of the developed software by calibration through a known size of the object. The software has been tested on many images of variety grains and results are found to be in agreement with the manual methods. In the case of length and width, the RMSE average values found to be 0.36 and 0.33, respectively. Keywords: Image processing, Size, Shape, Physical properties.