Notice of RetractionTemperature state recognition with simultaneous feature selection

Notice of Retraction

After careful and considered review of the content of this paper by a duly constituted expert committee, this paper has been found to be in violation of IEEE´s Publication Principles.

We hereby retract the content of this paper. Reasonable effort should be made to remove all past references to this paper.

The presenting author of this paper has the option to appeal this decision by contacting TPII@ieee.org.

Based on the complex fire reason analysis of metro vehicle, aiming at on-board high power devices of metro vehicle, e.g. , traction motor, the temperature state model of its surface and other related position is a most important index, which can reflects devices running normally or abnormally. Because there are lots of factors influencing or reflecting device temperature state model and neural network has special advantages in non-line fitting relationship between reasons and results, We apply self-organizing feature map (SOFM) neural network to recognize devices temperature state model through “recalling” the “memorized” weights for training and clustering past sample vectors. In order to improve its recognition accuracy and recognition speed, we take advantage of multi layer perception (MLP) network to extract useful and partially useful feature parameters, at the same time, eliminate bad feature parameters, these only extracted feature parameters are accepted by SOFM neural network. Through matlab simulation experiment, feature selection can diminish the dimensions of input vectors, only a very few parameters of one input vector, this network can satisfactorily achieve a good prediction of temperature state model.