A novel survival analysis of machine using fuzzy ensemble convolutional based optimal RNN

•SFEC-WSA algorithm is proposed to classify the device and identify survival time.•SFEC process both Sugeno fuzzy and the ABCDM to enhance the prediction results.•SFEC-WSA analyses parameters such as vibration, rotation, voltage and pressure. Survival Analysis is essential in the manufacturing field to determine unnecessary events by the input data. In Survival analysis, predictive maintenance plays a major portion in the identification of machine failures based on incoming input data from diverse equipment or sensors. Therefore, the Deep learning method is exploited for barbarizing the issues of predictive maintenance marginally but these techniques are not quite useful to predict the failure of devices for certain input data which the technique had not learned. Meanwhile, the neural network techniques are capable of predicting the output in accordance with the preceding input feature, the performance was poor when the input features have large variations. As a result, the transformation of input data degrades the performance of the neural network and the algorithm does not support the prediction of machine failure. To overcome such drawback, this paper proposes a novel Sugeno Fuzzy Ensemble Convolutional based War Strategy Algorithm (SFEC-WSA) to classify the device and identify the survival time in accordance with the input features. The proposed SFEC system integrates the process of both the Sugeno fuzzy integral ensemble model and the Attention-based Bidirectional CNN-RNN Deep Model (ABCDM). The SFEC-WSA algorithm is applicable in learning diverse input feature variations thereby predicting the robustness of the input data. The proposed SFEC-WSA analyses several parameters such as vibration, rotation, voltage, and pressure to evaluate the condition of the equipment. The experimentation results revealed that the proposed model effectively predicts large test data and performs better than other approaches.