Skip to content

Four-component NLP pipeline for essay content scoring and feedback generation

Design · All designs
Evidence · no claims cited

Description

A processing pipeline that takes learner-uploaded PDF essays and produces automated content feedback. The article describes "a pipeline composed of four components. These are 1) Preprocessing, 2) Segmentation, 3) Analytic scoring and 4) Feedback generation": text extraction and sentence tokenisation, sentence-level sliding-window classification into learning-tip sections, multilabel prediction of lecture-topic codes per segment against a ground truth table, and rule-triggered feedback texts displayed in a custom Moodle plugin. It was trained on manually labelled essays from a previous cohort and evaluated with a later cohort.

Design Implications

Context

Requirements

  • Manually labelled essays from a previous cohort as training data for both the segmentation and code-prediction steps
  • A ground truth table offered by the course instructors mapping correct concept assignments for each learning tip

Constraints

  • Implemented for one particular task: German-language essays about ten learning tips in an introductory educational psychology lecture

Target Learners

  • German-speaking university teacher students in an introductory educational psychology lecture

Learning Goals

  • Connecting learning tips to lecture concepts such as working memory, metacognition and motivation through argumentative essay writing

Claims

Examples

-

Key Sources

  • Gombert, S., Fink, A., Giorgashvili, T., Jivet, I., Di Mitri, D., Yau, J., Frey, A., & Drachsler, H. (2024). From the Automated Assessment of Student Essay Content to Highly Informative Feedback: a Case Study. International Journal of Artificial Intelligence in Education. https://doi.org/10.1007/s40593-023-00387-6
Page metadata
Field Value
type design
id essay-content-feedback-pipeline
title Four-component NLP pipeline for essay content scoring and feedback generation
description A processing pipeline that takes learner-uploaded PDF essays and produces automated content feedback.
status review
generated by process:settle-candidates on 2026-10-28
trust unverified — no one has confirmed this page's content
sources 1 — gombert-2024

Frontmatter is the page's machine-readable record — see the Schema & Guide.