基于Spark和Cloudera数据科学家工作平台的分布式机器学习实践.pdf

发布时间：2022-06-03 发布人：admin 分类：说明书资料大小：17.29M 资料格式：pdf 举报版权申诉

erjin_ren-10187543-4744302542971428348.pdf-第1页.png

第1页 / 共56页

erjin_ren-10187543-4744302542971428348.pdf-第2页.png

第2页 / 共56页

erjin_ren-10187543-4744302542971428348.pdf-第3页.png

第3页 / 共56页

erjin_ren-10187543-4744302542971428348.pdf-第4页.png

第4页 / 共56页

erjin_ren-10187543-4744302542971428348.pdf-第5页.png

第5页 / 共56页

erjin_ren-10187543-4744302542971428348.pdf-第6页.png

第6页 / 共56页

erjin_ren-10187543-4744302542971428348.pdf-第7页.png

第7页 / 共56页

erjin_ren-10187543-4744302542971428348.pdf-第8页.png

第8页 / 共56页

文本预览

数据科学改变传统行业 Connected Car Smart Industry Smart Cities & Ports Environment Sensing Usage Based Insurance Predictive Maintenance Aerospace & Aviation Smart Healthcare © Cloudera, Inc. All rights reserved. 2

Exploratory Data Science Required Capabilities: ● Unified Platform: ○Enables more workloads in a single platform at an integrated data science and engineering environment. ● Performance: ○Developers and data scientists have access to high scale and high performance query engines for distributed analytics. ● Ease of Use: ○Not for hacker users but data scientists. “Cloudera’s approach was more aligned with our own philosophies, such as building simpler, more prescriptive libraries that broaden the audience for the platform. ” “Cloudera Enterprise expedites round-trips to access and compute data for data discovery, translating into significant reductions in R&D time. This will have a very meaningful scientific upside.” © Cloudera, Inc. All rights reserved. 6

Machine Learning for Enterprise Required Capabilities: ●Expanded Data Access ○Expand your empirical data ●Test and Train Faster ○Iterative development ●Use Familiar Tools ○Increase developer productivity with familiar API’s ●Integrated Batch and Streaming ○unified batch and streaming programming model “Cloudera, using complex machine learning algorithms, analyzes large amounts of data in real time and allows personalization of game interaction with players through recommendations.,” “Machine learning and big data is like a marriage made in heaven. I mean it works really well with some other tooling that's already there on the stack. It used to take time. It was cumbersome.” “CDSW + Spark is perfect combination for machine learning on Hadoop. It simplified our work and save the cost of our data industry.” © Cloudera, Inc. All rights reserved. 7

Developing of Data Science 大量开源的机器学习框架 Data Science team(s) wants to use the latest open source tools IT/business don’t know how to support with existing analytics investments Hadoop平台 - 避免数据科学的”孤岛” - 数据供给和可视化加工问题 - 数据治理是数据科学的重要基础 -- 数据科学需要大量的计算资源，分布式资源调度、数据湖、私有云和多租户技术成为必然数据科学的相关组织架构 n 自助服务 – 数据服务、分析服务、建模服务 n 个性化的数据环境 n 业务部门和科技部门的价值平等，需要紧密协作 © Cloudera, Inc. All rights reserved. 8

分享到：

赞收藏

资料库

基于Spark和Cloudera数据科学家工作平台的分布式机器学习实践.pdf

相关推荐

大数据

热门标签

最新资料