Multiparadigm Data Science «多范式数据科学»

本文翻译自此网站

Multiparadigm Data Science is a new approach of using modern analytical techniques, automation and human-data interfaces to arrive at better answers with flexibility and scale. «多范式数据科学是一种利用现代分析技术、自动化和人-数据接口以具有灵活性和规模化的方式获得更好答案的新方法。»

Many organizations are still doing traditional data science—confining themselves to problems that are answerable with traditional statistical methods—rather than utilizing the broad range of interfaces and techniques available today. «许多组织仍在进行传统的数据科学研究,将自己局限于可以用传统统计方法解决的问题,而不是利用当今广泛的新界面和新技术。» Whether it’s automated machine learning, interactive notebooks and report generation, natural language queries of data for instant visualizations or implementing neural networks with ease and efficiency, modern problem solving requires access to the right technology at every stage. «无论是自动机器学习、交互式笔记本和报告生成、用于即时可视化的自然语言数据查询或轻松高效地实现神经网络,现代问题的解决都需要在每个阶段使用正确的技术。»

With a flexible, integrated multiparadigm workflow, problems too complex for traditional methods can get real, quantifiable answers. «通过灵活、集成的多范式工作流,对于传统方法来说过于复杂的问题可以得到真实、可量化的答案。»

Interfacing with Your Data «与你的数据连接»

Having the right interface to get answers from data is crucial; «拥有从数据中获得答案的正确界面至关重要;» different interfaces are suited to different tasks. «不同的交互界面适用于不同的任务。» A multiparadigm system combines a broad range of intuitive options for interacting with data every step of the way. «多范式系统结合了一系列直观的选项,可以在每一步与数据交互。»

Interactive Notebooks «交互式笔记本»

Streamline your daily workflow using high-level notebook documents that combine text, images, code and interactive examples—editable and runnable on any platform for fast, easy collaboration. «使用可在任何平台上编辑和运行的文本、图像、代码和交互式示例相结合的高级笔记本文档简化您的日常工作流程,实现快速、轻松的协作。»

Natural Language «自然语言»

Query your data instantly with natural language and get conversational answers for an intuitive, low-effort process that leads to more insights and better decisions. «使用自然语言即时查询数据,并获得对话式答案,以获得直观、低工作量的过程,获得更多的洞察力和更好的决策。»

Triggered Reporting «触发的报告»

Automatically receive alerts and updated reports based on custom criteria, tracking trends in real time for critically timed decisions. «根据自定义标准自动接收警报和更新的报告,实时跟踪趋势,以便进行关键的即时决策。»

Presentations with Live Modeling «带实时建模的演示文稿»

Run your meetings interactively, adjusting parameters to compute instant what-if scenarios for deeper discussions and stronger results. «以交互方式运行会议,调整参数以即时计算假设情景,以便进行更深入的讨论并获得更强大的结果。»

Programmatic Access «程序访问»

Store your models as platform-independent packages and APIs, providing a centralized framework for computation and automation to power human-data interfaces across your enterprise. «将您的模型存储为独立于平台的包和API,为计算和自动化提供一个集中的框架,以支持整个企业中的人-数据接口。»

Classic Interfaces with Modern Integration «经典界面与现代集成»

Seamlessly connect existing spreadsheets, databases and applications into higher-level computation, stepping up speed and accuracy with additional back end power. «无缝地将现有的电子表格、数据库和应用程序连接到更高级别的计算中,通过额外的后端功能提高速度和准确性。»

Know Your Data Science Areas «了解您的数据科学领域»

The field of data science is constantly growing and changing. «数据科学领域不断发展变化。» A multiparadigm workflow requires a broad algorithmic toolkit with the full suite of processing, analysis and visualization for ever-increasing computational needs. «一个多范式工作流需要一个广泛的算法工具箱,它具有全套处理、分析和可视化功能,以满足日益增长的计算需求。»

Machine Learning «机器学习»

Generate adaptive models directly from complex datasets for object classification and predictive analytics, such as identifying which new advertising markets to enter. «直接从复杂的数据集中生成自适应模型,用于对象分类和预测分析,例如确定要进入的新广告市场。»

Neural Networks «神经网络»

Create and train layered processing networks for deep analysis and processing tasks, such as recognizing defective items coming off a production line. «创建并培训分层处理网络,用于深入分析和处理任务,例如识别生产线中出现的缺陷项目。»

Dynamic Visualization «动态可视化»

Display data in styled plots, charts and infographics, making it human-readable and interactive for quick analysis and decision making. «以样式图、图表和信息图形显示数据,使其具有可读性和交互性,以便快速分析和决策。»

Data Semantics «数据感知»

Standardize various incoming datasets into a unified framework for easier analysis, such as consolidating data with different unit systems. «将各种输入数据集标准化为统一的框架,以便于分析,例如使用不同的单元系统整合数据。»

Systems Modeling «系统建模»

Model physical, electrical and other systems to inform design decisions, like the most effective heating installation for a building. «对物理、电气和其他系统进行建模,以告知设计决策,如建筑物最有效的加热装置。»

Optimization «优化»

Use high-level mathematics to discover the “best values” for your data in relation to key criteria, such as the ideal allocation of portfolio assets. «使用高级数学来发现与关键标准相关的数据的“最佳值”,例如投资组合资产的理想分配。»

Signal Processing «信号处理»

Process and filter images, audio and other collected data to analyze underlying patterns, such as detecting an irregular heartbeat from an ECG. «处理和过滤图像、音频和其他收集到的数据,以分析潜在模式,例如检测心电图中的不规则心跳。»

Geocomputation «地理计算学»

Use precise geolocation data and powerful geodetic computations to accurately examine real-world situations, such as visualizing optimal routes for a bus service. «使用精确的地理位置数据和强大的大地测量计算来精确地检查现实情况,例如可视化公交服务的最佳路线。»

Graph/Network Analysis «图/网络分析»

Explore and visualize systems of discrete relationships to analyze correlations and patterns, such as modeling demographics in a social network. «探索和可视化离散关系系统以分析相关性和模式,例如在社会网络中建模人口统计学。»

Cluster Analysis «聚类分析»

Group and analyze data based on similarity measures to extract underlying patterns and relationships, such as which customers are most similar to your top 100. «根据相似性度量对数据进行分组和分析,以提取基础模式和关系,例如哪些客户与您的前100名最相似。»

Survival Analysis «生存分析»

Compute survival functions and lifetime distributions to analyze time-to-event data, such as the expected lifetime of a piece of industrial equipment. «计算生存函数和生存期分布,以分析时间到事件数据,例如一件工业设备的预期寿命。»

Queueing Theory «排队论»

Model and simulate systems of queues to analyze waiting times and resource allocation, such as the optimal number of tellers at a bank branch. «对排队系统进行建模和模拟,以分析等待时间和资源分配,例如银行分支机构出纳员的最佳数量。»

Statistical Distributions «统计分布»

Fit historic data to parametric distributions to make inferences about the underlying events, such as the likelihood of a customer clicking through an ad. «将历史数据与参数分布相匹配,以对基础事件进行推断,例如客户点击广告的可能性。»

Morphological Analysis «形态分析»

Use geometric transformations on images and higher-dimensional data to analyze spatial properties, such as counting particles in a microscopic image. «对图像和更高维度的数据使用几何变换来分析空间属性,例如计算微观图像中的粒子数。»

Custom Interface Construction «自定义接口构造»

Make interactive onscreen controls for real-time adjustment of parameters in analyses and visualizations, allowing deeper exploration of data. «在屏幕上进行交互式控制,以便实时调整分析和可视化中的参数,从而更深入地探索数据。»

Mathematical Modeling «数学建模»

Drive systems of differential equations, recurrence relations and symbolic formulas with your data to test and refine models, such as computing the recovery rate of an epidemic. «驱动微分方程、递推关系和符号公式系统与您的数据一起测试和完善模型,例如计算一个流行病的恢复率。»

Time Series «时间序列»

Model, simulate and forecast sequences of events over time to track long-term trends and make predictions, such as expected sales for the next holiday season. «对一段时间内的事件序列进行建模、模拟和预测,以跟踪长期趋势并做出预测,例如下一个假日季的预期销售额。»

Semantic Text Analysis «语义文本分析»

Analyze underlying structures in linguistic data to clean up data and extract meaning, such as determining sentiment in customer comments. «分析语言数据中的底层结构,以清理数据并提取含义,例如确定客户评论中的情绪。»

Report Generation «报告生成»

Display conclusions and insights in a styled, formatted document for meetings, ongoing projects or public information, like a quarterly earnings report. «在样式化、格式化的文档中显示结论和见解,用于会议、正在进行的项目或公共信息,如季度收益报告。»

Wavelets «小波»

Deconstruct data signals into constituent parts for advanced manipulation and filtering of specific features, such as eliminating background noise from sensor data. «将数据信号分解为组成部分,以便对特定功能进行高级操作和过滤,例如从传感器数据中消除背景噪声。»

Random Processes «随机过程»

Model the progression of a system over time to make observations and predictions about its behavior, such as analyzing peak hours at a particular store location. «对系统随时间的发展进行建模,以对其行为进行观察和预测,例如分析特定商店位置的高峰时间。»

Computer Vision «计算机视觉»

Process visual data with machine learning and other sophisticated algorithms for analysis of features and patterns, such as identifying road hazards from a video feed. «使用机器学习和其他复杂算法处理视觉数据,以分析特征和模式,例如从视频源识别道路危险。»

Parallel Computing «并行计算»

Distribute parallel tasks to available computation units for large-scale scientific computing and other high-performance applications. «将并行任务分配给可用的计算单元,用于大规模科学计算和其他高性能应用。»

Thirty years of building the ultimate computation environment make the Wolfram technology stack ideal for Multiparadigm Data Science (MPDS)—a multitude of interfaces, language types, computational approaches and ready-to-use data all woven into one ecosystem. «30年的终极计算环境建设使Wolfram技术栈成为多范式数据科学(MPDS)的理想之选——许多接口、语言类型、计算方法和随时可用的数据都交织在一个生态系统中。»

A multiparadigm approach requires a broad, flexible computational toolkit that incorporates all aspects of a project into one start-to-finish workflow. «多范式方法需要一个广泛的、灵活的计算工具箱,它将项目的所有方面整合到一个从开始直到最终完成的完整工作流中。» The Wolfram technology stack does exactly this, enabling you to take data from hundreds of formats, carry out a full spectrum of analysis and visualization and immediately share or publish your results—all using the world’s largest collection of algorithms and computable knowledge. «Wolfram技术栈正是这样做的,使您能够从数百种格式中获取数据,进行全谱的分析和可视化,并立即使用世界上最大的算法和可计算知识集合共享或发布您的结果。»

广告

0 条回复 A 作者 M 管理员
    所有的伟大,都源于一个勇敢的开始!
欢迎您,新朋友,感谢参与互动!欢迎您 {{author}},您在本站有{{commentsCount}}条评论