GGLabYale/MTBench_finance_aligned_pairs_short

1---2dataset_info:3  features:4  - name: input_timestamps5    sequence: float646  - name: input_window7    sequence: float648  - name: output_timestamps9    sequence: float6410  - name: output_window11    sequence: float6412  - name: text13    dtype: string14  - name: trend15    dtype: string16  - name: technical17    dtype: string18  - name: alignment19    dtype: string20  splits:21  - name: train22    num_bytes: 13505850823    num_examples: 75024  download_size: 8792350025  dataset_size: 13505850826configs:27- config_name: default28  data_files:29  - split: train30    path: data/train-*31---32# MTBench: A Multimodal Time Series Benchmark33 34 35**MTBench** ([Huggingface](https://huggingface.co/collections/afeng/mtbench-682577471b93095c0613bbaa), [Github](https://github.com/Graph-and-Geometric-Learning/MTBench), [Arxiv](https://arxiv.org/pdf/2503.16858)) is a suite of multimodal datasets for evaluating large language models (LLMs) in temporal and cross-modal reasoning tasks across **finance** and **weather** domains.36 37Each benchmark instance aligns high-resolution time series (e.g., stock prices, weather data) with textual context (e.g., news articles, QA prompts), enabling research into temporally grounded and multimodal understanding.38 39## 🏦 Stock Time-Series and News Pair40 41This dataset contains aligned pairs of financial news articles and corresponding stock time-series data, designed to evaluate models on **event-driven financial reasoning** and **news-aware forecasting**.42 43### Pairing Process44 45Each pair is formed by matching a news article’s **publication timestamp** with a relevant stock’s **time-series window** surrounding the event46 47To assess the impact of the news, we compute the **average percentage price change** across input/output windows and label directional trends (e.g., `+2% ~ +4%`). A **semantic analysis** of the article is used to annotate the sentiment and topic, allowing us to compare narrative signals with actual market movement.48 49We observed that not all financial news accurately predicts future price direction. To quantify this, we annotate **alignment quality**, indicating whether the sentiment in the article **aligns with observed price trends**. Approximately **80% of the pairs** in the dataset show consistent alignment between news sentiment and trend direction.50 51 52###  Each pair includes:53 54- `"input_timestamps"` / `"output_timestamps"`: Aligned time ranges (5-minute resolution)55- `"input_window"` / `"output_window"`: Time-series data (OHLC, volume, VWAP, transactions)56- `"text"`: Article metadata57  - `content`, `timestamp_ms`, `published_utc`, `article_url`58  - Annotated `label_type`, `label_time`, `label_sentiment`59- `"trend"`: Ground truth price trend and bin labels60  - Percentage changes and directional bins (e.g., `"-2% ~ +2%"`)61- `"technical"`: Computed technical indicators62  - SMA, EMA, MACD, Bollinger Bands (for input, output, and overall windows)63- `"alignment"`: Label indicating semantic-trend consistency (e.g., `"consistent"`)64 65 66 67## 📦 Other MTBench Datasets68 69### 🔹 Finance Domain70 71- [`MTBench_finance_news`](https://huggingface.co/datasets/afeng/MTBench_finance_news)  72  20,000 articles with URL, timestamp, context, and labels73 74- [`MTBench_finance_stock`](https://huggingface.co/datasets/afeng/MTBench_finance_stock)  75  Time series of 2,993 stocks (2013–2023)76 77- [`MTBench_finance_aligned_pairs_short`](https://huggingface.co/datasets/afeng/MTBench_finance_aligned_pairs_short)  78  2,000 news–series pairs  79  - Input: 7 days @ 5-min  80  - Output: 1 day @ 5-min81 82- [`MTBench_finance_aligned_pairs_long`](https://huggingface.co/datasets/afeng/MTBench_finance_aligned_pairs_long)  83  2,000 news–series pairs  84  - Input: 30 days @ 1-hour  85  - Output: 7 days @ 1-hour86 87- [`MTBench_finance_QA_short`](https://huggingface.co/datasets/afeng/MTBench_finance_QA_short)  88  490 multiple-choice QA pairs  89  - Input: 7 days @ 5-min  90  - Output: 1 day @ 5-min91 92- [`MTBench_finance_QA_long`](https://huggingface.co/datasets/afeng/MTBench_finance_QA_long)  93  490 multiple-choice QA pairs  94  - Input: 30 days @ 1-hour  95  - Output: 7 days @ 1-hour96 97### 🔹 Weather Domain98 99- [`MTBench_weather_news`](https://huggingface.co/datasets/afeng/MTBench_weather_news)  100  Regional weather event descriptions101 102- [`MTBench_weather_temperature`](https://huggingface.co/datasets/afeng/MTBench_weather_temperature)  103  Meteorological time series from 50 U.S. stations104 105- [`MTBench_weather_aligned_pairs_short`](https://huggingface.co/datasets/afeng/MTBench_weather_aligned_pairs_short)  106  Short-range aligned weather text–series pairs107 108- [`MTBench_weather_aligned_pairs_long`](https://huggingface.co/datasets/afeng/MTBench_weather_aligned_pairs_long)  109  Long-range aligned weather text–series pairs110 111- [`MTBench_weather_QA_short`](https://huggingface.co/datasets/afeng/MTBench_weather_QA_short)  112  Short-horizon QA with aligned weather data113 114- [`MTBench_weather_QA_long`](https://huggingface.co/datasets/afeng/MTBench_weather_QA_long)  115  Long-horizon QA for temporal and contextual reasoning116 117 118 119## 🧠 Supported Tasks120 121MTBench supports a wide range of multimodal and temporal reasoning tasks, including:122 123- 📈 **News-aware time series forecasting**124- 📊 **Event-driven trend analysis**125- ❓ **Multimodal question answering (QA)**126- 🔄 **Text-to-series correlation analysis**127- 🧩 **Causal inference in financial and meteorological systems**128 129 130 131## 📄 Citation132 133If you use MTBench in your work, please cite:134 135```bibtex136@article{chen2025mtbench,137  title={MTBench: A Multimodal Time Series Benchmark for Temporal Reasoning and Question Answering},138  author={Chen, Jialin and Feng, Aosong and Zhao, Ziyu and Garza, Juan and Nurbek, Gaukhar and Qin, Cheng and Maatouk, Ali and Tassiulas, Leandros and Gao, Yifeng and Ying, Rex},139  journal={arXiv preprint arXiv:2503.16858},140  year={2025}141}142