Thumbnail Image

Chaudhary Charan Singh Haryana Agricultural University, Hisar

Chaudhary Charan Singh Haryana Agricultural University popularly known as HAU, is one of Asia's biggest agricultural universities, located at Hisar in the Indian state of Haryana. It is named after India's seventh Prime Minister, Chaudhary Charan Singh. It is a leader in agricultural research in India and contributed significantly to Green Revolution and White Revolution in India in the 1960s and 70s. It has a very large campus and has several research centres throughout the state. It won the Indian Council of Agricultural Research's Award for the Best Institute in 1997. HAU was initially a campus of Punjab Agricultural University, Ludhiana. After the formation of Haryana in 1966, it became an autonomous institution on February 2, 1970 through a Presidential Ordinance, later ratified as Haryana and Punjab Agricultural Universities Act, 1970, passed by the Lok Sabha on March 29, 1970. A. L. Fletcher, the first Vice-Chancellor of the university, was instrumental in its initial growth.

Browse

20171
Reset filters
Subject: Statistics × Author: Suman × End date: 2019 × Start date: 2010 ×
Reset filters

Search Results

Now showing 1 - 1 of 1
  • Thumbnail Image
    ThesisItemOpen Access
    ARIMA, state space and mixed modeling for sugarcane yield prediction in Haryana
    (CCSHAU, 2017) Suman; Verma, Urmil
    Forecasting of crop production is one of the most important aspects of agricultural statistics system. Crop production forecasting comprises crop identification, area estimation and predicting the yield of the crop. Understanding the behaviour of crop yields becomes increasingly important for modeling production functions, forecasting price movements and understanding the farmers’ responses to government programs. The statistical modeling approaches viz., ARIMA, state space and linear mixed modeling were used to achieve the district-level sugarcane yield estimation in major mustard growing districts of Haryana. The time-series sugarcane yield data for the period 1960-61 to 2009-10 of Karnal and Ambala districts, 1972-73 to 2009-10 of Kurukshetra district and 1980-81 to 2009-10 of Panipat and Yamunanagar districts were used for the development of different models. The selected models have been validated using the data on subsequent years i.e. 2010-11 to 2014-15, not included in the development of yield forecast models. After experimenting with different lags of moving average and autoregressive processes; ARIMA(0,1,1) for Karnal and Ambala districts and ARIMA(1,1,0) for Kurukshetra, Panipat and Yamunanagar districts were fitted. The underlying parameters of ARIMA models are assumed to be constant however the data in agriculture are generally collected over time and thus have the time-dependency in parameters. State space procedure giving time varying parameters models allow for known changes in the structure of the system over time. Thus, the same time series data were analyzed to achieve sugarcane yield estimates for the same five post-sample years using state space procedures by the application of Kalman filtering technique. Lastly, the linear mixed models with time both as fixed and random effects using different types of covariance structures viz., VC, AR(1) and Toeplitz were developed for sugarcane yield predition in the targeted districts. Finally, the performance of fitted models were decided on the basis of statistic(s) like AIC, BIC and log likelihood etc. Thus, the sugarcane yield estimates for the post-sample years 2010-11 to 2014-15 were obtained on the basis of fitted ARIMA, state space and linear mixed models. The predictive performance(s) of the contending models were observed in terms of percent deviations of sugarcane yield forecasts in relation to the observed yield(s) and root mean square error(s) as well. The state space models performed well with lower error metrics as compared to the alternative models in all time regimes i.e. these models consistently showed the superiority over ARIMA and linear mixed models in capturing percent relative deviations. In addition, the developed models are capable of providing the reliable yield estimates well in advance of the crop harvest while on the other hand, the DOA yield estimates are obtained quite late after the actual harvest of the crop.