CSIT696 - Research Methods in Computing

Investigation of Hyperspectral Data Applications

Montclair State University Spring 2021

Advisor Stefan Robila, PhD

Goals

• Learn about tools available to process document collections
  • Deliverable P2A
    • Processes a set of documents and finds the top 5 most frequently used words in the set (based on a given list of words).
  • Deliverable P2B
    • Processes a set of documents using some classic clustering algorithm.
• Machine grade ELA essays based on readability and vocabulary choice.
  • Deliverable P3B
    • Implementation of readability scores by Flesch-Kincaid, vocabulary scores by TF-IDF.
  • Deliverable P3C
    • Proof of concept on set of 15 ELA papers from a grade 9 classroom.
    • Machine grade each paper based on:
      • Readability Ratio
        • Each paper is assigned Coleman-Liau and Flesch-Kincaid readability score.
        • Score is divided by 12 for grade 12 being end of high school and assumption that an idealized paper should be at least in the 9-12 grade range.
      • Mean Vocabulary Score
        • Average TF-IDF score of 10 most frequent words in corpus for given paper.
        • Average TF-IDF score of 10 most frequend words in given paper.
      • Floor of Sum of Readability Ratio and Mean Vocabulary Score is multiplied by 100 and converted to integer in attempt to avoid inflating machine score.
        • Min Readability of Coleman-Liau and Flesch-Kincaid is used for avoiding inflation as well.