MeSCh: Measurement system for children's reading comprehension
Measurement System for Children's Reading Comprehension (MeSCh) is an automated multiple-choice test generation system that accepts a story text as input. The types of questions generated are knowledge (who, what, where and how), reasoning (why), sequencing and vocabulary. The target age for th...
Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | text |
| Language: | English |
| Published: |
Animo Repository
2008
|
| Subjects: | |
| Online Access: | https://animorepository.dlsu.edu.ph/etd_bachelors/14402 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Institution: | De La Salle University |
| Language: | English |
| Summary: | Measurement System for Children's Reading Comprehension (MeSCh) is an automated multiple-choice test generation system that accepts a story text as input. The types of questions generated are knowledge (who, what, where and how), reasoning (why), sequencing and vocabulary. The target age for the generated questions is six to nine years old. The story text is split into sentences, and then filtered to reduce to candidate sentences. Candidate terms are extracted and stems (questions) are formed from the candidate sentences. Two distractors (choices) are selected for each candidate term with the help of WordNet and database. Completed multiple-choice test items are stored in the system's question bank. Manual evaluation was performed by a child educator to validate the correctness of the stems and distractors generated by the system. A scale of 1 to 4 was used, where 1 reflects unusable stems or distractors, 2 and 3 for stems or distractors that can be used with major or minor modifications, respectively, and 4 for unusable stems or distractors. The stems underwent three evaluation phases, with each phase involving modifications to the algorithms to improve the quality of the output. Nine stories were used, and the stems received an average rating of 3.68% with 87.59% of the generated questions rated as usable in the third evaluation phase. This is due to missing of inappropriate definitearticles, complex or duplicate questions and inappropriate wh-word for the question. The distractors received an average rating of 3.47% with 68.59% of the generated distractors rated as usable. This is due to inappropriate choices. Missing or inappropriate definite articles, lack of punctuation marks and lack of relationship among the choices. |
|---|