The development of accurate real time disruption predictors is a pre-requisite to any mitigation action. 'Accurate real time prediction' has to be understood in terms of high success rate, low rates of false alarms as well as an early recognition of an incoming disruption. Only approximate theoretical models of disruptions exist and they do not reliably cope with the disruption issues. Therefore, this article deals with data-driven predictors with a view on ITER and DEMO. A review of existing prediction techniques, from both physics and engineering points of view, is provided. All these methods have to use large training datasets to develop successful predictors. However, ITER or DEMO cannot wait for hundreds of disruptions to have a reliable predictor. So far, the attempts to extrapolate predictors between different tokamaks have not shown satisfactory results. Moreover, it is not clear how valid this approach can be between present devices and ITER/ DEMO, due to the differences in their respective scales and possibly underlying physics. So, this article analyses the requirements to create adaptive predictors from scratch to learn from the data of an individual machine from the beginning of operation. A particular algorithm based on probabilistic classifiers has been developed and it has been applied to the database of the three first ITER-like wall experimental campaigns of JET (1036 non-disruptive and 201 disruptive discharges). The predictions show a success rate of 94%, a false alarm rate of 4.21% and an average warning time of 654ms. The average probability interval about the reliability and accuracy of all the individual predictions is 0.811 ± 0.189. It should also be mentioned that a very limited number of signals is required by the predictor, an important point particularly at the beginning of the operation of new devices.