Smartphone Sensor Dataset for Online Reading Analysis

Popularity of smartphones also popularized, reading content using smartphones. Reading using smartphones quite differs from reading using desktop system. Mouse and Keyboard are the peripherals associated with the reading in desktop systems. Study of the handling of such devices has led to provide implicit feedback of the content read. Similar study in smartphones to get implicit feedback remains to be a huge gap. Reading using smartphones involves screen gestures like pinch to zoom, tap, scroll, orientation change and screen capture. User reading behavior and intent can be done by screen gestures. Smartphones have sensors like gyroscope and accelerometer that continuously capture data without permissions from the users. To analyze screen gestures, a dataset has been created. Dataset is organized into accelerometer and gyroscope folders. This dataset can act as an input to train, test and validate, machine learning models to analyze smartphone reading behavior and user intent.

智能手机的普及也推动了利用智能手机进行内容阅读的行为流行开来。相较于桌面系统下的阅读场景，智能手机阅读存在显著差异。桌面系统下的阅读配套外设为鼠标与键盘。针对这类外设操作的相关研究，已实现通过阅读过程获取所阅读内容的隐式反馈。但针对智能手机端获取隐式反馈的同类研究仍存在巨大空白。智能手机阅读涉及多种屏幕手势操作，例如捏合缩放、轻点、滚动、屏幕方向切换与截屏。这类屏幕手势可用于推断用户的阅读行为与意图。智能手机内置陀螺仪（gyroscope）与加速度计（accelerometer）等传感器，可在无需用户授权的情况下持续采集数据。为开展屏幕手势分析相关研究，本数据集得以构建。该数据集按加速度计与陀螺仪两个文件夹进行分类存储。本数据集可作为训练、测试与验证机器学习模型的输入数据，用于分析智能手机阅读行为与用户意图。