It is observed that the spontaneous action potentials (sAPs) observed in a urinary bladder smooth muscle cell are of different shapes. The biophysical mechanisms underlying this shape variety is not yet known. It is assumed that the syncytial properties of the smooth muscle tissue in urinary bladder affect the shape of the sAPs generated by the smooth muscle cell. Further investigation on the matter requires accurate identification of different types of sAPs observable from the detrusor smooth muscle cells. Since such a ground truth on the number of possible sAP classes is not available and the manual identification of the sAP classes from long intracellular recording is tedious and erroneous, it becomes necessary to use an unsupervised classification algorithm to classify the observed sAP signals. K-means clustering and hierarchical clustering algorithms (both agglomerative and divisive approaches) are some of such classical clustering algorithms available. Also considering the different ways in which the data can be presented (such as raw time domain data, Fourier transform, wavelet transform, and principal components), There are multiple approaches to do the signal classification. In this study, the clustering results of all these approaches are compared and the best performing methods are shortlisted. An internal measure called cluster balance was used to quantitatively evaluate the resulting clusters.Mithun Padmakumar has completed his master in biomedical engineering from IIT Bombay in 2010 and bachelor in electrical and electronic engineering from College of Engineering Trivandrum, Kerala, India, in 2008.He is currently pursuing his Ph.D. degree in the Department of Biosciences and Bioengineering, Indian Institute of Technology (IIT) Bombay. He has attended the IBRO school on computational neuroscience in Hyderabad, and worked as a research associate in the Signal Processing and Instrumentation Lab, IIT Bombay where he developed algorithms for suppressing the motion artifacts and EMG noise from the ambulatory ECG recordings. His area of interest includes biomedical signal processing and pattern recognition.