The presentation will address the value of data that is "clean" or more accurately, of high quality versus "dirty" lower quality data. Issues and problems in assays that compromised data will be discussed as will the meta data and algorithms that can help to identify high vs low quality data. We will show examples of the results of using "kitchen sink" data vs carefully chosen data of high quality for generating predictive models. Additionally, the widely-forgotten distinct difference between machine learning and predictive modeling will be highlighted.

Cleansing drug discovery data for data science predictive modeling

255th National Meeting (2018)

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.