On 24 February 2022, Russia invaded Ukraine, starting what is now known as the Russo-Ukrainian War, initiating an online discourse on SNs. Twitter one of the most popular SNs, with an open and democratic character, enables a transparent discussion among its large user base. Unfortunately, this often leads to Twitter's policy violations, propaganda, abusive actions, civil integrity violations, and consequently to user accounts' suspension and deletion. This study focuses on the Twitter suspension mechanism and the analysis of shared content and features leading to an accurate machine-learning suspension prediction. Toward this goal, we have obtained a dataset containing 107.7M tweets, originating from 9.8 million users, using Twitter API. We extract the categories of shared content of the suspended accounts and explain their characteristics, through the extraction of text embeddings in junction with cosine similarity clustering. Our results reveal scam campaigns taking advantage of trending topics regarding the Russia-Ukrainian conflict for Bitcoin and Ethereum fraud, spam, and advertisement campaigns. Additionally, we apply a ML methodology including a SHapley Additive explainability model to understand and explain how user accounts get suspended.

This study proposes a three-stage, flexible and adaptable protocol for the establishment of field-scale agricultural management zones (AMZs) using remote sensing, ground truthing (apparent electrical conductivity and soil sampling), the IRRIGOPTIMAL^® system and machine learning. The methodology to develop this protocol was applied to olive and alfalfa plots in Heraklion (Crete, Greece) to monitor soil and plant responses for the period 2022–2024. However, the actual time for the implementation of this protocol varies between 3 and 6 months. The first step of this protocol involves the use of soil and vegetation reflectance mapping (moisture, photosynthetic activity) by satellites and unmanned aerial systems, together with geophysical electromagnetic induction mapping (apparent electrical conductivity) to verify soil variability, which is strongly linked to the delineation of management zones. In the second step, a machine learning-based prediction of the spatial distribution of soil electrical conductivity is made, considering the data obtained in the first step. Furthermore, in the second step, the IRRIGOPTIMAL^® system provides real-time monitoring of a variety of weather (such as air temperature, dew point, solar radiation, relative humidity, precipitation) and soil (temperature, moisture) parameters to support the optimal cultivation strategy for the plants. Once the data have been analysed, the soil variability of the plot and the presence or absence of cultivation zones are determined and the decision on the cultivation strategy is made based on targeted soil sampling and further soil analyses. This protocol could contribute significantly to the rational use of inputs (water, seeds, fertilizers and pesticides) and support variable rate technology in the agricultural sector of Crete.

This study proposes a flexible and adaptable protocol for the establishment of agricultural management zones that utilises remote sensing, ground truthing (apparent electrical conductivity and soil sampling), the IRRIGOPTIMAL® system and machine learning. This protocol could contribute significantly to the rational use of inputs (water, fertilizers and pesticides) and to the further efficient development of an optimal irrigation system with variable irrigation rates. The methodology to develop this protocol was applied to olive and alfalfa plots in Heraklion (Crete, Greece) to monitor soil and crop responses for the period 2022-2024. Spatial and temporal assessment of selected soil and plant parameters (moisture, photosynthetic activity) using ground and vegetation reflectance mapping by satellites and unmanned aerial systems provides important information in both the pre- and main phases of the management zone delineation. Geophysical methods such as electromagnetic induction, applied in the main phase of management zone delineation, provide a robust technique to determine the spatial and temporal distribution of apparent electrical conductivity. The Random Forest machine learning model was found to be most suitable for predicting soil electrical conductivity based on satellite-derived salinity indices and ground electromagnetic induction. Finally, the IRRIGOPTIMAL® system provides real-time monitoring of a variety of weather and soil parameters to determine the optimal cultivation of crops based on the creation of agricultural management zones.