Chronological horse herd optimization-based gene selection with deep learning towards survival prediction using PAN-Cancer gene-expression data

• Published in: Biomedical signal processing and control Vol. 84; p. 104696
• Main Authors: Majji, Ramachandro, Maram, Balajee, Rajeswari, R.
• Format: Journal Article
• Language: English
• Published: Elsevier Ltd 01.07.2023
• Subjects: Deep recurrent neural network; Gene expression data; Horse Herd Optimization; Political optimizer; Survival prediction; Survival prediction; Gene expression data; Political optimizer; Deep recurrent neural network; Horse Herd Optimization
• ISSN: 1746-8094; 1746-8108
• DOI: 10.1016/j.bspc.2023.104696

•The highlights of the article are given below for your kind perusal. Kindly, consider and forward my article for further processes.•A robust approach, named Deep Recurrent Neural Network-based Chronological Horse Herd Political Optimization (DRNN-based CHHPO) for survival prediction.•Here, the gene selection is performed using the proposed Chronological Horse Optimization (CHO) by considering the fitness parameters, such as Minkowski distance and Renyi entropy.•Meanwhile, the devised CHO is the hybridization of the Chronological concept and Horse Herd Optimization (HOA).•With the selected genes, the gene features are strengthened using technical indicators to enhance the overall process.•Finally, survival prediction is done using DRNN, which is trained by the designed optimization algorithm, named CHHPO, which is the incorporation of CHO and Political optimizer (PO).•The devised method achieved superior performance with the minimal Root Mean Square Error (RMSE), and Prediction Error (PE) of 0.467 and 0.456. Cancer has always been one of the major hazards to human life which is also the most difficult part of human disease history. The death rate due to cancer is high. The prediction results are affected because of the major dissimilarities present in clinical results. Hence, it is necessary to enhance the accuracy of cancer survival prediction, which remains a challenging one. To defeat the challenges, this research devises a robust approach, named Deep Recurrent Neural Network-based Chronological Horse Herd Political Optimization (DRNN-based CHHPO) for survival prediction. Here, the gene selection is performed using the proposed Chronological Horse Optimization (CHO) by assuming the parameters of fitness, for example Minkowski distance plus Renyi entropy. The Horse Herd Optimization (HOA) and Chronological concept is merged to form the CHO. With the selected genes, the gene features are strengthened using technical indicators to enhance the overall process. Finally, survival prediction is completed by means of DRNN, which is trained by the CHHPO, which is the amalgamation of Political optimizer (PO) and CHO. Superior presentation with the Prediction Error (PE) and minimal Root Mean Square Error (RMSE) of 0.456 and 0.467 is accomplished by this developed technique.