<p>Liver diseases are a significant global health concern affecting millions worldwide. Prevention, early diagnosis and treatment are crucial in mitigating complications and improving patient quality of life. Machine learning techniques are essential for supporting medical decisions and improving diagnosis precision. In this context, this work presents a collaborative system based on a features selection voting technique that identifies the most important feature to define the disease presence or absence. The most representative features identified through this voting process were employed in the Fuzzy <i>c</i>-means clustering algorithm. Furthermore, a ranking method was developed, which proved to be a valuable tool for measuring the patient risk and analyze anomalous values. In summary, the proposed method demonstrated substantial advantages in a clinical setting, providing diverse information that supports clinical decision-making in terms of liver disease diagnosis.</p>