Calad/fake-TGC

---
license: apache-2.0
---

# Dataset Card

## Dataset Details

This dataset is primarily created for the work [Fast muon tracking with machine learning implemented in FPGA](http://dx.doi.org/10.1016/j.nima.2022.167546) ([Arxiv link](https://arxiv.org/abs/2202.04976)) that contains ~3M simulated muon events with Geant4.
Hits in the muon chamber and ground truth of track angle are saved.

### Dataset Description

Please refer to the Section Simulation samples in the referenced work for details.

The file contains 3 keys: 'X', 'Y', and 'corr'.

'X' is a boolean array of size (3072000, 7, 50) used as the input information.

'Y' is a float vector of size (3072000) that contains the ground truth angle to be predicted.

'corr' contains three keys each of size (100, 100) that contains the Pearson correlation factor between the named stations that can be derived from X. Not useful for general purpose.

## Uses

```python
import h5py as h5
with open('dataset.h5','r') as f:
X = np.array(f['X'])
Y = np.array(f['Y'])
```

## Dataset Structure

```
<ROOT>
├── X: bool[3072000, 7, 50]
│
├── Y: float64[3072000]
│
└── corr
├── 12: float64[100, 100]
│
├── 23: float64[100, 100]
│
└── 13: float64[100, 100]
```

## Citation [optional]

You can cite the original work that introduces this dataset.

**BibTeX:**

```
@article{Sun_2023,
title={Fast muon tracking with machine learning implemented in FPGA},
volume={1045},
ISSN={0168-9002},
url={http://dx.doi.org/10.1016/j.nima.2022.167546},
DOI={10.1016/j.nima.2022.167546},
journal={Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment},
publisher={Elsevier BV},
author={Sun, Chang and Nakajima, Takumi and Mitsumori, Yuki and Horii, Yasuyuki and Tomoto, Makoto},
year={2023},
month=jan, pages={167546}
}
```

原始信息汇总

数据集概述

数据集详情

• 创建目的: 用于支持研究论文《Fast muon tracking with machine learning implemented in FPGA》，该论文描述了在FPGA中实现机器学习进行快速μ子追踪的方法。
• 数据内容: 包含约300万次模拟μ子事件，使用Geant4模拟。数据集记录了μ子室中的击中情况及轨迹角度的真实值。

数据集结构

• X: 布尔数组，大小为(3072000, 7, 50)，作为输入信息。
• Y: 浮点数向量，大小为(3072000)，包含待预测的真实轨迹角度。
• corr: 包含三个键，每个键的大小为(100, 100)，存储了各站之间的皮尔逊相关系数，不用于一般用途。

数据集使用

• 数据集存储于.h5文件中，可通过Python的h5py库读取。

数据集引用

• 可通过引用原始论文《Fast muon tracking with machine learning implemented in FPGA》来引用此数据集。

• BibTeX引用:

  @article{Sun_2023, title={Fast muon tracking with machine learning implemented in FPGA}, volume={1045}, ISSN={0168-9002}, url={http://dx.doi.org/10.1016/j.nima.2022.167546}, DOI={10.1016/j.nima.2022.167546}, journal={Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment}, publisher={Elsevier BV}, author={Sun, Chang and Nakajima, Takumi and Mitsumori, Yuki and Horii, Yasuyuki and Tomoto, Makoto}, year={2023}, month=jan, pages={167546} }

搜集汇总

背景与挑战

背景概述

该数据集包含约300万个模拟μ子事件，使用Geant4生成，主要用于机器学习在FPGA中实现快速μ子跟踪的研究。数据集中包括布尔型输入数组X、浮点型真实角度Y以及相关性数据corr，支持相关学术论文的复现和实验。