Pulmonary function analysis is essential for assessing respiratory health, with spirometry serving as the standard method. However, its continuous use is limited by the need for specialized equipment. Inductive-bands provide a portable alternative, measuring thoracic expansion, yet converting their raw signals into respiratory-volume estimates requires additional processing. In this study, machine-learning models were employed to transform the inductive-band signal into respiratory-volume signals derived from spirometric measurements. A total of 1,180 recordings were collected from healthy volunteers across four respiratory tests. The acquired signals underwent filtering, differentiation, integration, and temporal alignment. Subsequently, a diverse set of machine-learning methods was trained and evaluated using cross-validation. Results showed that a convolutional neural network (CNN) achieved the best performance. The findings revealed a high correlation between the estimated respiratory volume and the scaled inductive-band signal, and the machine-learning models further refined the estimated spirometry signals, supporting the validity of the proposed approach. As future work, the CNN model will be optimized to improve its accuracy for high-frequency components, and embedded-hardware implementations will be explored for real-time respiratory-monitoring applications.