Monday (8/2) Response (Abstract)

(1) Presentation Title:

• Predicting Student Success

(2) Abstract First Draft

• I have chosen to pursue trying to create a model that can help determine how likely a student will be to need additional assistance or tutoring based on their characteristics. To explore this I found a dataset on Kaggle that has information on 2,133 different students. The target variable within this dataset is the student’s end-of-year post-test score. The dataset also contains ten different features, ranging from the type of school the student is enrolled in, to information specific to the student. This information can be used to train a model that can predict a student’s post-test score, which will allow school systems to have the ability to predict how well each student will score and have the opportunity to try and give them extra help. I will train the model using feature columns to train a continuous target model. I’ll then experiment with different combinations of the various features to create the best model possible.

• Breakdown of Data

  • 2,133 Students within Dataset

  • 10 Features:

    • School

    • School Setting: Urban, Rural, or Suburban

    • School Type: Public or Non-public

    • Classroom: I’m unsure has so what these features mean

    • Teaching Method: Standard or Experimental

    • Number of Students in Class

    • Student ID

    • Gender: Male or Female

    • Lunch: Does not qualify or Qualifies for reduced/free lunch

    • Pretest: Score student got on pretest assessment

  • Target is Student Post-Test Score

    • Range of 32% to 100%