The phytoplankton community plays a key role in ecosystem functioning. Monitoring their biodiversity and dynamics is essential for assessing ecosystem responses to diverse environmental drivers, including but not limited to climate change. Traditional phytoplankton analysis relies on inverted light microscopy (Utermöhl method), an accurate but time-consuming and resource-intensive technique. Moreover, typical sampling frequencies are not always adequate for detecting rapid ecological events, such as harmful algal blooms (HABs). To overcome these limitations, oceanographic research is increasingly adopting automated imaging instruments and machine learning algorithms. Among these technologies, the Imaging FlowCytobot (IFCB) combines flow cytometry with high-resolution imaging, accelerating data acquisition rates and enabling high-frequency observations of phytoplankton communities. However, the large volume of images generated requires training algorithmic workflows to automate taxonomic classification. 
This study focuses on the implementation, standardization, and preliminary evaluation of Random Forest classifier developed using IFCB images collected at the Promontory of Portofino eLTER (Long-Term Ecological Research) site (Ligurian Sea). Classifier performance was evaluated using precision, recall, and F1-score metrics.  
Although further optimization is needed, the classifier showed promising performance, with F1-scores near to 1 for morphologically distinct and homogeneous groups, whereas it becomes weaker for classes characterized by high internal morphological variability or for the rare taxa. This limitation is primarily due to the insufficient number of images available to adequately cover and represent all taxonomic categories. Future developments will focus on expanding and balancing the training dataset, improving classifier performance for underrepresented taxa, exploring the use of Convolutional Neural Networks (CNNs), and integrating the workflow into collaborative image classification platforms such as EcoTaxa.