Design and Application of Scientific Research Management System based on Time Loss function with Recurrent Neural Network

The scientific research management system is a comprehensive software solution that is used to manage research activities of solvent scientific organizations. This research proposes a Time Loss function with Recurrent Neural Network (TLF-RNN) for scientific research management system. The TLF is used to reduce the implication of time delay or absences of analyzing data points which is usual in scientific research programs. The incorporation of TLF-RNN in a scientific research management system enhanced ability to capture temporal changes, increased predictive accuracy in order to optimize their management and planning in scientific research. The min-max scalar is used as preprocessing which transforms all the observation within the given dataset in order to bring the minimum and maximum value of an observation to the range of 0 and 1. Then, the features are extracted by Term Frequency-Inverse Document Frequency (TFIDF) which is applied for quantify term weight identified in a specific document. Moreover, it increases predictive accuracy of the model by providing better distinction on the documents. The TLF-RNN attains precision 0.942, accuracy 0.956, f-measure 0.931 and recall 0.947 which is better than other existing techniques.