Sustainable agriculture is currently being challenged under climate change scenarios since extreme environmental processes disrupt and diminish global food production. For example, drought-induced increases in plant diseases and rainfall caused a decrease in food production. Machine Learning and Agricultural Big Data are high-performance computing technologies that allow analyzing a large amount of data to understand agricultural production. Machine Learning and Agricultural Big Data are high-performance computing technologies that allow the processing and analysis of large amounts of heterogeneous data for which intelligent IT and high-resolution remote sensing techniques are required. However, the selection of ML algorithms depends on the types of data to be used. Therefore, agricultural scientists need to understand the data and the sources from which they are derived. These data can be structured, such as temperature and humidity data, which are usually numerical (e.g., float); semi-structured, such as those from spreadsheets and information repositories, since these data types are not previously defined and are stored in No-SQL databases; and unstructured, such as those from files such as PDF, TIFF, and satellite images, since they have not been processed and therefore are not stored in any database but in repositories (e.g., Hadoop). This study provides insight into the data types used in Agricultural Big Data along with their main challenges and trends. It analyzes 43 papers selected through the protocol proposed by Kitchenham and Charters and validated with the PRISMA criteria. It was found that the primary data sources are Databases, Sensors, Cameras, GPS, and Remote Sensing, which capture data stored in Platforms such as Hadoop, Cloud Computing, and Google Earth Engine. In the future, Data Lakes will allow for data integration across different platforms, as they provide representation models of other data types and the relationships between them, improving the quality of the data to be integrated.

In the context of Big Data scenarios, the presence of extensive static datasets is not uncommon. To facilitate efficient queries on such datasets, the utilization of multiple indexes, such as the Kd-tree, becomes imperative. The current scale of managed points may, however, exceed the capacity of primary memory, posing a significant challenge. In this article we introduce cKd-tree, a compact data structure designed to represent a Kd-tree efficiently. The structure cKd-tree is essentially an encoding of the spiral code sequence of points within an implicit Kd-tree (iKd-tree) using Directly Addressable Codes (DACs). The unique feature of cKd-tree lies in its ability to perform spiral encoding and decoding of points by relying solely on knowledge of their parent points within the iKd-tree. This inherent property, combined with DACs’ direct access capability to sequence elements, enables cKd-tree to traverse and explore the tree while decoding only the nodes relevant to queries. The article details the algorithms necessary for creating and manipulating a cKd-tree, as well as algorithms for evaluating two fundamental queries over points: the point query and the range query . To assess the performance of cKd-tree, a series of experiments are conducted, comparing it with iKd-tree and k ² -tree data structures. The evaluation metrics include compression efficiency and execution time of queries. cKd-tree achieves a compression ratio comparable to that of k ² -tree, approximately 70%, demonstrating heightened efficiency, particularly in scenarios characterized by sparse data. Additionally, consistent with expectations, k ² -tree exhibits superior performance in querying individual points, whereas cKd-tree outperforms in the context of aggregate data queries, such as range queries.

Proximity searches within metric spaces are critical for numerous real-world applications, including pattern recognition, multimedia information retrieval, and spatial data analysis, among others.With the exponential increase in data volume, the demand for memory-efficient structures to store and process information has become increasingly important. In this paper, we present an alternative algorithm for efficient computation of the K nearest neighbors (KNN) query using the k²-tree compact data structure, using the incremental radius technique. This approach offers an alternative to the existing algorithm that utilizes a priority queue over k₂-trees. Through both theoretical and experimental analysis, we demonstrate that our proposed algorithm is up to 2 times faster compared to the priority queue-based solution, while also providing substantial improvements in memory efficiency.