ºÝºÝߣ

ºÝºÝߣShare a Scribd company logo

Feature Engineering for Second Language Acquisition Modeling

?
?
Guanliang Chen
Guanliang Chen

The poster for my research on knowledge tracing in the 11th Workshop on Innovative Use of NLP for Building Educational Applications (co-located with NAACL HLT) June 5 in New Orleans.

1 of 1
Download to read offline
Feature Engineering for Second Language Acquisition Modeling Guanliang Chen, Claudia Hauff, Geert-Jan Houben Web Information Systems, TU Delft English Spanish French High school Primary school Mathematical learning Motivation1 Research Question2 What factors impact students' language learning performance? Knowledge tracing Second language acquisition ¡­ Invested time Learning environment Prior knowledge Step 2: Gradient Tree Boosting 4 Knowledge Tracing Step 1: Feature Engineering Gradient Tree Boosting is most effective with 9 of the designed features, e.g., learning type and exercise format, for second language acquisition modeling. More efforts on feature engineering are needed. 3 Research Hypotheses TU Delft Extension School A student¡¯s living community affects her learning performance.H1 The more engaged a student is, the more words she masters.H2 The more time a student spends on solving an exercise, the more likely she will get it wrong. H3 Contextual factors such as the device being used, learning type and exercise format impact a student¡¯s learning. H4 Repetition is useful and necessary for a student to master a word. H5 A student with a high-spacing learning routine is more likely to learn more words than one with a low-spacing learning routine.?H6 Analyzing by grouping students living in different countries. Analyzing by grouping students according to their spent time and learning spacing routine. 23 features

More Related Content

Feature Engineering for Second Language Acquisition Modeling

  • 1. Feature Engineering for Second Language Acquisition Modeling Guanliang Chen, Claudia Hauff, Geert-Jan Houben Web Information Systems, TU Delft English Spanish French High school Primary school Mathematical learning Motivation1 Research Question2 What factors impact students' language learning performance? Knowledge tracing Second language acquisition ¡­ Invested time Learning environment Prior knowledge Step 2: Gradient Tree Boosting 4 Knowledge Tracing Step 1: Feature Engineering Gradient Tree Boosting is most effective with 9 of the designed features, e.g., learning type and exercise format, for second language acquisition modeling. More efforts on feature engineering are needed. 3 Research Hypotheses TU Delft Extension School A student¡¯s living community affects her learning performance.H1 The more engaged a student is, the more words she masters.H2 The more time a student spends on solving an exercise, the more likely she will get it wrong. H3 Contextual factors such as the device being used, learning type and exercise format impact a student¡¯s learning. H4 Repetition is useful and necessary for a student to master a word. H5 A student with a high-spacing learning routine is more likely to learn more words than one with a low-spacing learning routine.?H6 Analyzing by grouping students living in different countries. Analyzing by grouping students according to their spent time and learning spacing routine. 23 features