Non-intrusive loading monitoring (NILM) provides a smart solution to the problem of electrical energy monitoring of households at appliance level. In blind disaggregation, the power level of each appliance is not known a priori. In this paper we propose an event based blind online disaggregation algorithm which uses Gaussian mixture models (GMM) for clustering to automatically detect two state appliances from the aggregate data. The benefit of using a Gaussian mixture models over other clustering methods is that they are able to learn automatically the statistical distributions present in the data. This is beneficial, especially when we have appliances that have similar power consumption. Since Gaussian mixture models do not determine the number of clusters automatically, we use Bayesian information criteria (BIC) to determine the number of clusters. The blind disaggregation method is tested with data from a real house collected by a smart meter which samples the aggregate consumption at 3.4 kHz and also from Reference Energy Disaggregation Dataset (REDD) public data, sampled at frequency of 1 Hz. We saw improved performance by using Gaussian mixture models instead of mean shift clustering in various accuracy measures. We were able to track the total power consumed 12 % better then by mean shift clustering; root mean square error is also considerably lower on our measured data. The performance of the algorithm on the public data is comparable to other methods.