Prediction of major pest incidence in Jute crop based on weather variables using statistical and machine learning models: A case study from West Bengal
| Autor: | null PRAHLAD SARKAR, null PRADIP BASAK, null CHINMAYA SUBHRAJYOTI PANDA, null DEB SANKAR GUPTA, null MRINMOY RAY, null SABYASACHI MITRA |
|---|---|
| Rok vydání: | 2023 |
| Předmět: | |
| Zdroj: | Journal of Agrometeorology. 25:305-311 |
| ISSN: | 2583-2980 0972-1665 |
| DOI: | 10.54386/jam.v25i2.1915 |
| Popis: | Jute crop cultivated in Cooch Behar suffers a substantial amount of physical and economical loss every year due to several major insect pest infestation such as Yellow Mite (Polyphagotarsonemus latus Banks) and Jute Semilooper (Anomis sabulifera Guen). Constructed seasonal plots reveal that for Yellow Mite pest incidence is maximum at 55 DAS, while for Jute Semi Looper it is at 45 DAS. Correlation analysis indicate that the weather parameters such as minimum temperature at current week, maximum RH at one week lag, minimum temperature, minimum and maximum RH at two week lag are significantly correlated with the incidence of Yellow Mite, while in case of Jute Semilooper maximum temperature, minimum and maximum RH at two week lag are significantly correlated. Different forecasting models like ARIMA, ARIMAX, SARIMA, SARIMAX and SVR have been fitted and validated using RMSE values. In case of Jute Semilooper, SARIMAX model is found to be the best fitted model followed by SVR and SARIMA. Similarly, for Yellow Mite ARIMAX model produces the least RMSE value followed by SVR and ARIMA. Following successful model validation, forecasting is done for the year 2022 using the best fitted models. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|