Abstract: Data Science practitioners often come across datasets that where data-points from a particular class are rarer than the others. This talk discusses strategies that one can incorporate to deal with class-imbalances for classification.
Brief Description of the Content: The talk provides a description of class-imbalances and then delves into methodologies that go into improving classification performance on such datasets. Further, a project on Fraud Detection is described to illustrate the applications of the discussed methodologies.
Prerequisites: Experience with simple-classification.
Abstract: Data Science practitioners often come across datasets that where data-points from a particular class are rarer than the others. This talk discusses strategies that one can incorporate to deal with class-imbalances for classification.
Brief Description of the Content: The talk provides a description of class-imbalances and then delves into methodologies that go into improving classification performance on such datasets. Further, a project on Fraud Detection is described to illustrate the applications of the discussed methodologies.
Prerequisites: Experience with simple-classification.
Time Required: ~30 minutes
Link to Slides: https://bit.ly/ImbLearnPyData
Will you be doing a hands-on demo as well? No
Link to iPython Notebook: http://bit.ly/2TT1wbe
About Yourself: I love Data Science. Other things I love are street food and rock music.
Are you comfortable if the talk is recorded and uploaded to PyData Delhi's YouTube channel? Yes
Any query?