فهرست نشریات علمی دسترسی آزاد ایران

Similarity measures play a critical role in quantifying relationships between data points. Recent advancements have expanded the exploration of these measures into neutrosophic sets, a framework capable of simultaneously addressing uncertainty, indeterminacy, and inconsistency in datasets. Among these, cosine similarity measures for neutrosophic sets have attracted significant attention for their ability to improve pattern recognition accuracy, particularly in classifying complex or ambiguous patterns. Instead of using classical cosine similarity measures in vector spaces, applying neutrosophic truth, indeterminacy, and falsity membership values enables broader applications in multi-attribute decision-making. Jaccard similarity measures focus on the ratio of intersecting sets to the union of sets, while improved Jaccard measures for neutrosophic sets further address indeterminacy, making them particularly effective for clustering, document matching, and evaluating overlaps between ideal solutions. This study highlights the uncertainty and complexity of the data by proposing a hybrid Cosine-Jaccard similarity measure for neutrosophic sets. Some key properties, such as symmetry, boundedness, and reflexivity, are carefully examined to ensure their consistency and reliability. The practical examples demonstrate the measure’s validity by emphasizing its potential to advance decision-making and pattern recognition.