A Statistical Framework for Predictive Model Evaluation in MOOCs. Gardner, J. & Brooks, C. In Proceedings of the Fourth (2017) ACM Conference on Learning @ Scale, of L@S '17, pages 269–272, Cambridge, MA, USA, April, 2017. Association for Computing Machinery. Paper doi abstract bibtex Feature extraction and model selection are two essential processes when building predictive models of student success. In this work we describe and demonstrate a statistical approach to both tasks, comparing five modeling techniques (a lasso penalized logistic regression model, naïve Bayes, random forest, SVM, and classification tree) across three sets of features (week-only, summed, and appended). We conduct this comparison on a dataset compiled from 30 total offerings of five different MOOCs run on the Coursera platform. Through the use of the Friedman test with a corresponding post-hoc Nemenyi test, we present comparative performance results for several classifiers across the three different feature extraction methods, demonstrating a rigorous inferential process intended to guide future analyses of student success systems.
@inproceedings{gardner_statistical_2017,
address = {Cambridge, MA, USA},
series = {L@{S} '17},
title = {A {Statistical} {Framework} for {Predictive} {Model} {Evaluation} in {MOOCs}},
isbn = {978-1-4503-4450-0},
url = {http://doi.org/10.1145/3051457.3054002},
doi = {10.1145/3051457.3054002},
abstract = {Feature extraction and model selection are two essential processes when building predictive models of student success. In this work we describe and demonstrate a statistical approach to both tasks, comparing five modeling techniques (a lasso penalized logistic regression model, naïve Bayes, random forest, SVM, and classification tree) across three sets of features (week-only, summed, and appended). We conduct this comparison on a dataset compiled from 30 total offerings of five different MOOCs run on the Coursera platform. Through the use of the Friedman test with a corresponding post-hoc Nemenyi test, we present comparative performance results for several classifiers across the three different feature extraction methods, demonstrating a rigorous inferential process intended to guide future analyses of student success systems.},
urldate = {2020-09-23},
booktitle = {Proceedings of the {Fourth} (2017) {ACM} {Conference} on {Learning} @ {Scale}},
publisher = {Association for Computing Machinery},
author = {Gardner, Josh and Brooks, Christopher},
month = apr,
year = {2017},
keywords = {machine learning, model evaluation, mooc, predictive modeling},
pages = {269--272}
}
Downloads: 0
{"_id":"6afZhYrxaipKmcEfT","bibbaseid":"gardner-brooks-astatisticalframeworkforpredictivemodelevaluationinmoocs-2017","authorIDs":[],"author_short":["Gardner, J.","Brooks, C."],"bibdata":{"bibtype":"inproceedings","type":"inproceedings","address":"Cambridge, MA, USA","series":"L@S '17","title":"A Statistical Framework for Predictive Model Evaluation in MOOCs","isbn":"978-1-4503-4450-0","url":"http://doi.org/10.1145/3051457.3054002","doi":"10.1145/3051457.3054002","abstract":"Feature extraction and model selection are two essential processes when building predictive models of student success. In this work we describe and demonstrate a statistical approach to both tasks, comparing five modeling techniques (a lasso penalized logistic regression model, naïve Bayes, random forest, SVM, and classification tree) across three sets of features (week-only, summed, and appended). We conduct this comparison on a dataset compiled from 30 total offerings of five different MOOCs run on the Coursera platform. Through the use of the Friedman test with a corresponding post-hoc Nemenyi test, we present comparative performance results for several classifiers across the three different feature extraction methods, demonstrating a rigorous inferential process intended to guide future analyses of student success systems.","urldate":"2020-09-23","booktitle":"Proceedings of the Fourth (2017) ACM Conference on Learning @ Scale","publisher":"Association for Computing Machinery","author":[{"propositions":[],"lastnames":["Gardner"],"firstnames":["Josh"],"suffixes":[]},{"propositions":[],"lastnames":["Brooks"],"firstnames":["Christopher"],"suffixes":[]}],"month":"April","year":"2017","keywords":"machine learning, model evaluation, mooc, predictive modeling","pages":"269–272","bibtex":"@inproceedings{gardner_statistical_2017,\n\taddress = {Cambridge, MA, USA},\n\tseries = {L@{S} '17},\n\ttitle = {A {Statistical} {Framework} for {Predictive} {Model} {Evaluation} in {MOOCs}},\n\tisbn = {978-1-4503-4450-0},\n\turl = {http://doi.org/10.1145/3051457.3054002},\n\tdoi = {10.1145/3051457.3054002},\n\tabstract = {Feature extraction and model selection are two essential processes when building predictive models of student success. In this work we describe and demonstrate a statistical approach to both tasks, comparing five modeling techniques (a lasso penalized logistic regression model, naïve Bayes, random forest, SVM, and classification tree) across three sets of features (week-only, summed, and appended). We conduct this comparison on a dataset compiled from 30 total offerings of five different MOOCs run on the Coursera platform. Through the use of the Friedman test with a corresponding post-hoc Nemenyi test, we present comparative performance results for several classifiers across the three different feature extraction methods, demonstrating a rigorous inferential process intended to guide future analyses of student success systems.},\n\turldate = {2020-09-23},\n\tbooktitle = {Proceedings of the {Fourth} (2017) {ACM} {Conference} on {Learning} @ {Scale}},\n\tpublisher = {Association for Computing Machinery},\n\tauthor = {Gardner, Josh and Brooks, Christopher},\n\tmonth = apr,\n\tyear = {2017},\n\tkeywords = {machine learning, model evaluation, mooc, predictive modeling},\n\tpages = {269--272}\n}\n\n","author_short":["Gardner, J.","Brooks, C."],"key":"gardner_statistical_2017","id":"gardner_statistical_2017","bibbaseid":"gardner-brooks-astatisticalframeworkforpredictivemodelevaluationinmoocs-2017","role":"author","urls":{"Paper":"http://doi.org/10.1145/3051457.3054002"},"keyword":["machine learning","model evaluation","mooc","predictive modeling"],"downloads":0},"bibtype":"inproceedings","biburl":"https://api.zotero.org/groups/2554149/items?key=LIJbpMnFOlMFWOhvB0IgbufR&format=bibtex&limit=100","creationDate":"2020-10-09T20:55:54.181Z","downloads":0,"keywords":["machine learning","model evaluation","mooc","predictive modeling"],"search_terms":["statistical","framework","predictive","model","evaluation","moocs","gardner","brooks"],"title":"A Statistical Framework for Predictive Model Evaluation in MOOCs","year":2017,"dataSources":["BxTs4dXArXHzHDKT6"]}