A Brief Overview on Change-Point Methods
Change-point methods (CPMs) are statistical techniques meant to assess whether a given sequence of i.i.d. stochastic observations can be separated in two (or more) consecutive parts, each generated by a different distribution [1]. The earliest works in this direction operate on data generated from a known statistical distribution (parametric CPM), however, both nonparametric and online CPM meant to sequentially analyze data streams have been recently proposed [2]. 
The presentation aims at introducing the core mechanism of parametric and offline CPM, as well as the most recent extensions (nonparametric and online CPM). As a final remark the I will present an ensemble of CPM that is specifically meant for data that are not i.i.d., since the temporal dependencies in the sequence may impair CPM performance [3].

References
[1] D. M. Hawkins and P. Qiu, “The changepoint model for statistical process control,” JOURNAL OF QUALITY TECHNOLOGY, vol. Vol.35, No. 4, pp. 355–366, 2003.
[2] G. J. Ross, D. K. Tasoulis, and N. M. Adams, “Nonparametric monitoring of data streams for changes in location and scale,” Technometrics, vol. 53, no. 4, pp. 379–389, 2011.
[3] C. Alippi, G. Boracchi, V. Puig, M. Roveri, "An Ensemble Approach to Estimate the Fault-Time Instant," in IEEE Proceedings of ICICIP 2013, International Conference on Intelligent Control and Information Processing, June 9 - 11, Beijing, China -- To Appear