Meteorological Time Series Clustering in Agricultural Applications: A Systematic Literature Review

Context: Clustering of meteorological time series in the agricultural context is extremely useful for improving agricultural decision support systems, mainly through climate zoning. Problem: Given the particularities of meteorological time series, the clustering task is complex, usually involving data preprocessing and feature extraction steps, in addition to the need to keep up with the advancement of technology and machine learning techniques. Solution: This study brings together the main solutions for clustering meteorological time series in agricultural applications, in a context-aware way, mapping the main challenges and seeking to understand the characteristics of the meteorological data, in order to better understand the applicability of different techniques in the agricultural context. IS Theory: This work was developed within the scope of Argumentation Theory, gathering and compiling data from primary studies on the topic, as well as evidence that proves the legitimacy of these data and conclusions in the form of statements. Method: This study presents a descriptive and systematic literature review, according to a well-defined and widely used methodology, regarding published works on clustering of meteorological time series in agricultural applications. Summary of Results: After an initial search, the papers were screened and filtered based on the review protocol, and 26 papers were selected for review. Data were then extracted about the solutions presented in each paper, such as objective, operation, experiments, and evaluation metrics. IS Contributions and Impact: The main contribution of this study is the organization of published knowledge on the research topic, in order to identify the state-of-the-art and assist researchers, as well as the discussion and highlighting of future research directions.