A web app which generates the automated summary from different algorithms, and compares them with manual summaries.
Make a web app, which has an interface which lets you take a textbook pdf as input, and gives the automated summary of the different algorithms in different sections of the page, and a manual summary alongside it. Then compare each of the outputs of the automated summary with the manual version by developing an algorithm which checks the context in which each sentence comes in, and hence the meaning of the sentence, rather than parsing it and checking it sentence by sentence, and output the accuracy of it, or infer something from it.
Steps: 1)Implement these algorithms and generate the automated summary given a textbook pdf. 2)Generate a web app interface which lets you take an input, and does the above step 3)Make an interface where the user can enter his manual summary 4)Compare the two summaries with an algorithm(Still to be designed) 5)Infer.
Tools: 1)Microsoft Cognitive Services(MCS) 2)Python: i)scikit-learn library; ii)nltk iii)numPy