Chapter 15 - Machine learning approach for climate change impact assessment in agricultural production

Abstract

Agricultural production mainly depends on weather conditions. Both aspects are interiorly connected with each other in several aspects, as climate change is the key factor of plant biotic and abiotic stresses, which have an adverse influence on global agriculture production. The agricultural land is being affected by climate changes in several ways, for example, variations in annual rainfall and temperature, weed modifications, microbial activity, heat waves, and global atmospheric CO2, or ozone level change. The warning of global climate change has significantly driven the attention of researchers, as these alterations are imparting adverse impacts on crop production and negotiating the global food security system. A timely accurate forecast of crop production is essential for critical policy assessments such as pricing, export-import, marketing circulation, and global food security. The emerging empirical model relationship is being widely used to evaluate the effects of climate change on a specific region. The problematic situation is deriving information from raw datasets, this has led to the expansion of new techniques such as machine learning (ML) that can be used significantly to integrate the information with crop yield assessment. This chapter is designed as an attempt to present crop yield modeling that uses ML methods in high-dimensional datasets. Some statistical approaches such as artificial neural networks, decision tree modeling, regression analysis, fuzzy networks, Markov chain modeling, principal component analysis, cluster analysis, and time-series analysis were summarized.