Numerous scalable approaches are developed for mining various types of patterns, with frequent pattern mining being a key theme in data mining research. The internet is getting considerably more complicated, which makes it both more crucial and more difficult to process different data mining challenges across diverse areas. This research propose novel technique in frequent pattern data mining based on sequence using clustering and DL architecture. Frequent pattern based data is clustered and classified using fuzzy clustering with back propagation ResNet Convolutional networks. The experimental analysis has been carried out in terms of accuracy, precision, recall, F-1 score and RMSE. Proposed technique attained accuracy of 95%, precision of 71%, recall of 65%, F-1 score of 77% and RMSE of 62%.