A Framework for Extractive Text Summarization Based on Deep Learning Modified Neural Network Classifier

On account of the exponential augmentation of documents on the internet, users need all the pertinent data at ?1? place with no hassle. Therefore, automatic text summarization (ATS) is needed to automate the procedure of summarizing text via extorting the salient details as of the documents. The goal is to propose an automatic, generic, in addition to extractive text summarization for a single document utilizing Deep Learning Modifier Neural Network (DLMNN) classifier for generating an adequately informative summary centered upon the entropy values. A proposed DLMNN framework comprises ?6? phases. In the initial phase, the input document is pre-processed which engages stop word removal, tokenization, along with stemming. Subsequently, the features are extorted as of the pre-processed data. Next, the most apposite features are selected employing the improved fruit fly optimization algorithm (IFFOA). The entropy value for every chosen feature is computed utilizing support as well as confident measure. Afterward, DLMNN classifier is utilized to classify these values into ?2? classes, a) highest entropy values and b) lowest entropy values. Lastly, the class that holds the highest entropy values are chosen besides, the informative sentences are selected as of the highest entropy values to form the last summary. Experimental outcomes are executed and the proposed DLMNN classifier?s performance is analyzed utilizing sensitivity, accuracy, recall, specificity, precision, and also f-measure. The proposed DLMNN provides the best outcomes amid all other techniques.