SERVICES FOR CREATING A TOOLKIT FOR LEARNING AND PREDICTING ALCOHOL USE PATTERNS.

用于创建学习和预测酒精使用模式的工具包的服务。

• 批准号: 10272783
• 负责人: JAMES BAURLEY, PH.D.
• 金额: $ 22.45万
• 依托单位: BIOREALM
• 依托单位国家: 美国
• 财政年份: 2020
• 资助国家: 美国
• 起止时间: 2020-09-08 至 2021-03-07
• 项目状态: 已结题
• 来源: https://reporter.nih.gov/project-details/10272783
• 关键词: Alcohol consumption; Alcoholic beverage heavy drinker; Alcohols; Algorithms; Biological; Contractor; Contracts; Data; Learning; Longitudinal Studies; Modeling; National Institute on Alcohol Abuse and Alcoholism; Patients; Pattern; Phase; Research Personnel; Risk; Services; Small Business Innovation Research Grant; alcohol research; innovation; machine learning algorithm; multidimensional data; prototype; simulation; support vector machine

This contract is for a project titled, "Toolkit for Learning and Predicting Alcohol Use Patterns (SBIR Phase I, Topic 019), and the purpose for this project is for creating a platform to accelerate alcohol research that consists of developing an innovative toolkit for learning and predicting alcohol use patterns and creating machine learning algorithms that are capable of analyzing high dimensional data, which shall have the ability to model the more complicated features inherent in biological data. In this Phase I project, the Contractor shall demonstrate the feasibility of the proposed platform by extending the support vector machines (SVMs) to handle random effects, profiling the algorithms through simulations, and applying the prototype platform to an NIAAA-sponsored longitudinal study of heavy drinkers. The data, algorithms, and results shall be packaged into interactive Jupyter notebooks for alcohol researchers to explore and compute risk in new patients.

这份合同是为一个项目，题为“学习和预测酒精使用模式的工具包（SBIR第一阶段，主题 019），该项目的目的是创建一个平台，以加速酒精研究，包括 开发用于学习和预测酒精使用模式的创新工具包，并创建机器学习 能够分析高维数据的算法，这些算法应该能够对更多的数据进行建模。 生物数据中固有的复杂特征。在本一期工程中，承包商应论证其可行性 通过扩展支持向量机（SVM）来处理随机效应，分析 算法通过模拟，并应用原型平台NIAAA赞助的纵向研究， 酗酒者应将数据、算法和结果打包到酒精的交互式自动分析仪笔记本中 研究人员探索和计算新患者的风险。