Abstract:
This study aims to model and predict student academic performance in the university-level
course Vector Algebra and Geometry using Lagrange polynomial interpolation. This
research was conducted in a controlled academic environment where teaching and syllabus
variables remained constant. Historical data from 2015 to 2021 were analysed using
MATLAB software for computation and visualization. Data from 2015–2019 were used for
interpolation; 2017 was tested for interpolation accuracy, while 2020 and 2021 served as
extrapolation cases. Key performance indicators, including Average Marks, Maximum and
Minimum Marks, and also Pass Rates were assessed. The results affirm that Lagrange
interpolation offers meaningful insights into student outcomes when applied within the data
range, with moderate limitations in extrapolation. The model showed high accuracy within
known data ranges but reduced performance during extrapolation. However, extrapolated
predictions showed substantial deviations, with errors exceeding 100%, indicating that the
model performs poorly beyond the known data range. The findings demonstrate the potential
of mathematical modeling for educational data analysis and forecasting.
