2 个仓库
Organizing tabular data into structured dataframes for analysis.
Distinct from Data Structures: Distinct from general Data Structures: focuses specifically on tabular dataframe implementations for data science.
Explore 2 awesome GitHub repositories matching data & databases · Dataframe Structures. Refine with filters or upvote what's useful.
statsforecast 是一个高性能统计时间序列预测库,旨在生成点预测和预测区间。它作为一个分布式时间序列框架,利用基于 C 的预测引擎和自动模型选择器来识别并拟合数据集中每个唯一序列的最佳统计模型。该系统还包括一个时间序列异常检测器,通过将观测值与概率预测区间进行比较来识别异常数据点。 该项目的特色在于其处理数百万个独立序列的大规模并行预测的能力。它通过分布式计算框架、多核并行执行和加速核心 ARIMA 及指数平滑逻辑的编译 C 内核来实现这一点。该系统进一步利用长格式数据布局和惰性求值数据流水线来优化大规模处理,以减少内存开销。 该库提供了一套全面的模型,包括 AutoARIMA、用于间歇性或季节性需求的各种指数平滑方法、Theta 分解以及用于金融风险的 GARCH 波动率建模。它涵盖了更广泛的功能领域,例如带有外生变量的多元预测、时间序列分解以及通过历史交叉验证和滑动窗口分析进行模型评估。 该库与 Polars 等高性能数据结构集成,并提供将保存的模型作为 REST 端点提供服务以进行网络可访问预测的实用程序。
Employs high-performance dataframe structures to organize multiple time series for efficient memory management.
tqsdk-python is a quantitative trading SDK and framework designed for developing automated strategies for futures, options, and stocks using Python. It functions as an algorithmic trading engine and financial market data API, providing the tools necessary to backtest strategies, analyze historical data, and execute live trades across multiple brokerage accounts. The project distinguishes itself through a specialized option analytics library that calculates Greeks, implied volatility, and volatility surfaces using the Black-Scholes model. It further supports complex order execution patterns, s
Converts sequential market tick and k-line data into structured DataFrames for optimized numerical analysis.