基于睡眠传感器的高频鼾声片段占比数据

以睡眠传感器采集到的睡眠数据为基础，高频鼾声片段占比有以下使用场景。1）个人健康管理：专业人员可以通过分析睡眠数据中的高频鼾声占比,即时评估鼾声严重程度，并反馈用户，该数据的比例值越高表示打鼾症状越严重。字段说明：
1）夜间每10秒鼾声检测数据表示每个10秒的信号片段是否发生了打鼾事件，0表示未发生打鼾事件，1表示发生了打鼾事件。
2）夜间每10秒所有频段能量和表示每个10秒的信号片段转换为频域特征后，所有频段的能量和。能量和为频域特征的平方和。
3）夜间每10秒所有频段能量和表示每个10秒的信号片段转换为频域特征后，频段大于180hz的能量和。

算法概要：
1. 针对夜间每10秒所有频段能量和以及夜间每10秒高频段能量和字段中缺失的数据，使用上一个10秒片段的数据进行填充。
2. 根据夜间每10秒鼾声检测数据字段，筛选出发生打鼾事件的所有片段序列，即选出每10秒鼾声检测数据字段值为1的片段。
3. 针对筛选出来的所有片段，将这些片段对应的夜间每10秒所有频段能量和字段数据与夜间每10秒所有频段能量和字段数据求比值并求和，得到的数值即为高频鼾声片段占比字段数据。

Based on sleep data collected by sleep sensors, the proportion of high-frequency snoring segments has the following application scenarios.
1) Personal health management: Professionals can analyze the proportion of high-frequency snoring in sleep data to instantly evaluate the severity of snoring and provide feedback to users. A higher proportion value corresponds to more severe snoring symptoms. Field descriptions:
1. Nighttime 10-second snoring detection data: Indicates whether a snoring event occurred in each 10-second signal segment. A value of 0 denotes no snoring event, while a value of 1 denotes occurrence of a snoring event.
2. Nighttime 10-second total energy across all frequency bands: Represents the sum of energy across all frequency bands after converting each 10-second signal segment into frequency-domain features. The total energy is calculated as the sum of squares of the frequency-domain features.
3. Nighttime 10-second high-frequency energy sum (frequency bands above 180Hz): Represents the sum of energy of frequency bands above 180Hz after converting each 10-second signal segment into frequency-domain features.
Algorithm Overview:
1. For missing values in the fields of nighttime 10-second total energy across all frequency bands and nighttime 10-second high-frequency energy sum (above 180Hz), fill them using the data from the preceding 10-second segment.
2. Filter all segment sequences where snoring events occurred based on the nighttime 10-second snoring detection data field, i.e., select segments with a value of 1 in the 10-second snoring detection data field.
3. For all the selected segments, compute the ratio of their corresponding nighttime 10-second high-frequency energy sum (above 180Hz) data to their corresponding nighttime 10-second total energy across all frequency bands data, then sum these ratios. The resulting value is the data for the high-frequency snoring segment proportion field.

搜集汇总

背景与挑战

背景概述

该数据集提供基于睡眠传感器采集的高频鼾声片段占比数据，采用CSV格式存储，包含1个文件约1MB大小，适用于睡眠呼吸障碍相关研究。