Autoregressive integrated moving average (ARIMA) technique has been widely used for forecasting in divergent domains for several decades. One of the main drawbacks of this model is the presumption of linearity. As many time series phenomenon in real word are nonlinear hence, it is required to enhance the prediction ability of ARIMA models by amalgamating it with other nonlinear models. One of the sound techniques to achieve this goal is to combine linear and nonlinear time series models so that forecasting efficacy of the models can be improved. In the present research paper, an attempt is made to combine the ARIMA model with Time delay neural network (TDNN) and also with nonlinear support vector regression (NLSVR) model. As a case study, banana production in Karnataka has been considered to evaluate the forecasting performance of the hybrid models under study. Empirical results clearly reveal that the forecasting accuracy of the hybrid models are better as compared to the ARIMA model. Keywords: Banana production, ARIMA, TDNN, NLSVR, Hybrid methodology.

  The Odumade and Singh (2009) randomized response (RR) technique involving two decks of cards is used to estimate the proportion of individuals belonging to a certain sensitive group. In this paper, Odumade and Singh (2009) RR technique has been extended to k-decks of cards. The proposed alternative estimator for k-decks cards is more efficient than the existing Odumade and Singh (2009) estimator for k = 2. The proposed variance estimator of the RR technique is simple and nonnegative. Keywords: Complex sampling designs; Randomized response; Relative efficiency.

  In some cases, occurrence of binary data may vary in different spatio-temporal situations. The logistic regression fails to model binary data from clustered, multi-level and longitudinal studies, because of dependency among the observations. Under such situation, random effects can be included in the linear predictor of the logistic regression model in order to allow for correlated responses. The estimation of parameter of the binary logistic regression with random effects (LRRE) is not straight forward due to the fact that the likelihood involves multiple integrals and the explicit derivation of the integrals is not possible. The Bayesian paradigm provides a natural approach to inference in mixed models. In this paper we modeled the occurrence of aphid populations in two different locations in India using logistic regression with random effects in Bayesian paradigm. We consider that the response variable is binary in nature. The weather variables i. e. temperature, relative humidity and their interaction are taken as covariates. We assume different prior distributions for the random intercepts parameters. As we know odds ratio depends on random intercepts, so it is also shown that odds ratio is random. Hence median odds ratio is calculated and it is shown that median odds ratio is robust and better measure than odds ratio. SAS software version 9.4 has been used for present analysis. Keywords: Aphid population, Bayesian paradigm, LRRE model, Random intercept.

  Most common methods of estimation where the auxiliary information are used ratio, regression and product methods of estimation. Calibration estimation has developed into an important field of research in survey sampling where the auxiliary information plays an important role. In the present paper an efforts has been made to develop calibration estimator in under two-stage stratified random sampling design when auxiliary information relative to variable under study are available at element level for selected primary stage unit (psu) i.e auxiliary information is available at secondary stage unit (ssu) for selected primary stage unit (psu). A simulation study has been conducted to investigate the relative performance of calibration estimator over the usual estimator of the population mean without using auxiliary information in two-stage stratified random sampling. Keywords: Calibration approach, Calibration estimator, Auxiliary information, Two-stage sampling, Stratified random sampling.

  In the present article an improved estimator for estimating finite population variance under measurement errors has been proposed. Bias and mean squared error of proposed estimator have been obtained up to first order of approximation. Theoretical efficiency comparison has been made among proposed estimator, sample variance estimator, usual ratio estimator and estimators proposed by Misra et al. (2016). A numerical illustration has been made using hypothetical data generated through R software. Interpretation of results is also shown through graphical representation of mean squared errors of estimators with and without measurement errors. Keywords: Auxiliary information, Bias, Efficiency, Measurement errors, Mean squared error, Population variance, Simple random sampling, R Software.

  Fisheries sector in India has made giant strides in recent years and its role in increasing food supply, generating job opportunities, raising nutritional level and earning foreign exchange need hardly be emphasized. Its future is likely to be much different from what it is now. New upcoming technologies and resources are expected to be different from what these are now which would require reconciliations of conflicting socio-economic and environmental objectives and trade-offs. Significant development in fisheries has been made possible by technological progress and policy changes. However, in the context of increasing globalization and emergence of new markets, the future of Indian fisheries is bound to be influenced. It is therefore imperative to articulate technological needs of different aspects of fishing and fish processing and contemplate how developments in science and policy can help to address these needs. For this, applications of two chief tools from the field of Technology Forecasting (TF) have been employed in the domain of fisheries. Of the TF tools available, two techniques viz., Cross Impact Analysis (CIA) technique and Substitution modeling were attempted in the field of fisheries. One of CIA techniques viz., Kane?s simulation model was employed to study the long-term dynamics of production (marine), imports and exports of Indian fish commodities. In addition, by considering global production of P. vannamei shrimp species and other shrimp species over time, substitution models viz., Pearl and Gompertz were fitted to ascertain the time period required for the emerging species to replace the traditional ones. The analogy between the CIA model used and substitution models are also explained. To sum up, it has been demonstrated that TF can give incisive insights that can aid decision makers to plan strategies in transforming expected outcomes to preferred futures. Keywords: Kane?s simulation modeling, Gompertz model, Pearl model, Foresight, vannamei shrimp.

  This paper examined the transmission mechanism of an increase in energy prices on the price of selected agricultural commodities in India using monthly wholesale price indices during April 1994 to March 2014. In order to assess the effect of deregulation of some petroleum products which was implemented since April, 2002, the sample data was divided into two periods (April 1994 - March 2004 and April 2004 - March 2014), besides analyzing for full period so that before and after period analysis will provide clearer picture of possible link between the two prices. To supplement the finding of Johansen co-integration analysis, we assessed the nature and extent of causal relationship between the variables. The results supplement the finding of Johansen co-integration analysis, we assessed the nature and extent of causal relationship between the variables. The results ndicated evidence of parallel movement between prices of energy and all selected agricultural commodities after deregulation, which means higher transmission between crude oil and commodities prices. The results obtained are expected to help in understanding of transmission mechanism for policy makers to optimize and stabilize the markets. Keywords: Agricultural commodity prices, Energy, Co-integration.

  In this paper, we generalize the Herfindahl-Hirschman Index (HHI) using the correlation structure of the individual shares to arrive at a new diversity measure. This is then applied to the cultivation of crops across zones of Assam, India. Before applying these indices we have removed the trend component from the data series. Since the trend effect is eliminated, the indices show the true diversity pattern and reflect the intentions of farmers directly which was not taken into consideration in the earlier measures. The results show that the diversity indices declined till 1975-76 and then gradually increased up to 1987-88. Thereafter, all the indices became more or less stagnant. It may be because the farmers did not want to take risk or they have less access to the modern farming technology which is also subject to changes over time. Keywords: Crop diversity, Herfindahl-Hirschman index, Hall and Tideman index, Cropping pattern of Assam

  N.A

  N.A

  Indian Council of Agricultural Research (ICAR) under National Agricultural Research and Education System (NARES) carries out training programs popularly known as Capacity Building Program (CBP) for its vast manpower engaged in teaching, research and extension. To support in training management, efficiency and knowledge sharing an e-governance initiative was taken up by ICAR and a web based system was developed and implemented (http://cbp.icar.gov.in). The system was designed in modular approach with built in support for document integration and in-built work flows for identified users (Course Directors, Participants and Managers). The work flows in system deals with various stages of training management starting with proposal submission, approval, financials, participant application to feedback submission and training report upload at the end of the training program. System also provides a platform for knowledge sharing in the form of e-books which is spontaneously generated from the lecture notes uploaded for the training programs. The information base of the system enables the managers in taking policy level decisions for training planning, management and consolidating the future requirements of the human resources. At present, the system has a databank of 12570 users, 1190 training proposals and about 225 e-books (e-resources) which will go long way with the system usage. Keywords: Human Resource Development, Indian Council of Agricultural Research, National Agricultural Research and Education System, Training Management, e-books in Agriculture.