Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

DeepDyve requires Javascript to function. Please enable Javascript on your browser to continue.

Auto-Grader - Auto-Grading Free Text Answers

Part of the BestMasters Book Series

DOI: 10.1007/978-3-658-39203-1

ISBN: 978-3-658-39202-4

Publisher: Springer Fachmedien Wiesbaden - Springer Journals

Oct 15, 2022

LitStream

Chapter 1Introduction

Richner, Robin

Pages: 1 –2

[As follows the research problem, objective and anticipated contribution are stated.]
LitStream

Chapter 2Research Design

Richner, Robin

Pages: 3 –4

[The master thesis follows the Cross-Industry Standard Process for Data Mining (CRISP-DM). CRISP-DM has six high level phases: business understanding, data understanding, data preparation, modelling, evaluation and deployment. However, the deployment will not be part of the thesis.]
LitStream

Chapter 3Research Background

Richner, Robin

Pages: 5 –26

[Firstly, this chapter will introduce the technological background needed to understand how a state-of-the-art auto-grader may look and secondly elaborate on related work in the field of automatic grading.]
LitStream

Chapter 4Data

Richner, Robin

Pages: 27 –44

[This chapter commences with a description of the provided data, continues with an analysis and preprocessing of the data and ends with a preprocessed data set in the form of a pandas dataframe as a base for the model development. Note that there will still be adjustments to the data depending on...
LitStream

Chapter 5Model Development

Richner, Robin

Pages: 45 –62

[There are five models described as follows which can be divided into three categories. The models in the categories differ in their data augmentation strategy and in their architecture. Data augmentation refers to how the data is fed into the model. The architecture refers to the specific neural...
LitStream

Chapter 6Evaluation

Richner, Robin

Pages: 63 –73

[This chapter has the following structure. Firstly, the models described in chapter 5 are trained and tested on a small data set due to computational limitations as previously discussed. Secondly, the two best-performing models are trained on all data. Thirdly, the hyperparameters are tuned. More...
LitStream

Chapter 7Discussion, limitations and further research

Richner, Robin

Pages: 75 –83

[This chapter discusses the thesis, elaborates on the limitations and further research opportunities. Hereby, it is divided into preprocessing, data augmentation, pre-training, fine-tuning, and bias.]
LitStream

Chapter 8Conclusion

Richner, Robin

Pages: 85 –86

[The thesis introduced the time constraint that teachers face when it comes to grading free text answer questions. The objective was to create a system that would assist teachers to save time on that task. It could be seen that related work consists of various AI and non AI-based approaches and...

results per page

Prev
1
Next

Browse All eBooks