Said Bahassine

Part-of-Speech (POS) tagging is a foundational task in Natural Language Processing (NLP), yet it remains challenging for low-resource and morphologically rich languages such as Amazigh. This paper proposes a hybrid ensemble framework for Amazigh POS tagging that integrates three complementary models: a Bidirectional Long Short-Term Memory network (BiLSTM), a Conditional Random Field model (CRF), and a rule-based morphological analyzer (RBMA). Rather than treating prior results obtained on different corpora and tag inventories as directly comparable, the study evaluates all proposed components under a common 54-tag experimental setting based on the publicly available Amazigh Linguistic Dataset. Three ensemble strategies are examined: majority voting, validation-weighted voting, and logistic-regression stacking. An additional late-fusion ablation applies hard and soft RBMA constraints to CRF and Stacking outputs; hard masking degrades performance substantially, whereas soft masking is more robust but remains below unconstrained decoding. The best micro-level performance is obtained by the stacking ensemble, which reaches 98.51% Micro-F1/accuracy, whereas the boosting-like weighted ensemble obtains the strongest Macro-F1 among the ensemble variants, reaching 74.24%. These results show that hybrid ensemble methods can improve token-level accuracy in low-resource POS tagging, while also revealing a trade-off between frequent-tag accuracy and rare-tag robustness. The findings highlight the usefulness of combining neural, probabilistic, and rule-based information for Amazigh POS tagging, and point to class-balanced meta-learning and character/subword representations as important directions for improving rare and out-of-vocabulary categories.

Feature selection (FS) is a widely used method for removing redundant or irrelevant features to improve classification accuracy and decrease the model’s computational cost. In this paper, we present an improved method (referred to hereafter as RARF) for Arabic text classification (ATC) that employs the term frequency-inverse document frequency (TF-IDF) and Word2Vec embedding technique to identify words that have a particular semantic relationship. In addition, we have compared our method with four benchmark FS methods namely principal component analysis (PCA), linear discriminant analysis (LDA), chi-square, and mutual information (MI). Support vector machine (SVM), k-nearest neighbors (K-NN), and naive Bayes (NB) are three machine learning based algorithms used in this work. Two different Arabic datasets are utilized to perform a comparative analysis of these algorithms. This paper also evaluates the efficiency of our method for ATC on the basis of performance metrics viz accuracy, precision, recall, and F-measure. Results revealed that the highest accuracy achieved for the SVM classifier applied to the Khaleej-2004 Arabic dataset with 94.75%, while the same classifier recorded an accuracy of 94.01% for the Watan-2004 Arabic dataset.

Over the course of the previous two decades, there has been a rise in the quantity of text documents stored digitally. The ability to organize and categorize those documents in an automated mechanism, is known as text categorization which is used to classify them into a set of predefined categories so they may be preserved and sorted more efficiently. Identifying appropriate structures, architectures, and methods for text classification presents a challenge for researchers. This is due to the significant impact this concept has on content management, contextual search, opinion mining, product review analysis, spam filtering, and text sentiment mining. This study analyzes the generic categorization strategy and examines supervised machine learning approaches and their ability to comprehend complex models and nonlinear data interactions. Among these methods are k-nearest neighbors (KNN), support vector machine (SVM), and ensemble learning algorithms employing various evaluation techniques. Thereafter, an evaluation is conducted on the constraints of every technique and how they can be applied to real-life situations.