Spoken Arabic Digits

Dataset from 8800(10 digits x 10 repetitions x 88 speakers) time series of 13 Frequency Cepstral Coefficients (MFCCs) had taken from 44 males and 44 females Arabic native speakers between the ages 18 and 40 to represent ten spoken Arabic digit.

This is a pre-processed version of the dataset saved in numpy format. The original dataset is obtained from UCI.

The data are 3-dimensional arrays of shape [n_samples, time_steps, n_variables]. The data can be loaded as follows:

loaded_data = np.load("ARAB.npz")
Xtr = loaded_data['Xtr'] # Training data of shape (6600, 93, 13)
Ytr = loaded_data['Ytr'] # Training labels of shape (6600, 1)
Xte = loaded_data['Xte'] # Test data of shape (2200, 93, 13)
Yte = loaded_data['Yte'] # Test labels of shape (2200, 1)

本数据集包含8800组13维梅尔频率倒谱系数（Mel-Frequency Cepstral Coefficients, MFCCs）时序序列，采集自年龄介于18至40岁的88名母语为阿拉伯语的发音人（男性44名、女性44名）录制的10个阿拉伯语口语数字：每位发音人对每个数字重复录制10次，总样本量为10个数字×10次重复×88名发音人，总计8800组。

本数据集为预处理后的版本，以NumPy格式存储。原始数据集源自UCI。

该数据集为三维数组，形状格式为[n_samples, time_steps, n_variables]。数据可通过以下代码加载：
loaded_data = np.load("ARAB.npz")
Xtr = loaded_data['Xtr'] # 形状为(6600, 93, 13)的训练数据
Ytr = loaded_data['Ytr'] # 形状为(6600, 1)的训练标签
Xte = loaded_data['Xte'] # 形状为(2200, 93, 13)的测试数据
Yte = loaded_data['Yte'] # 形状为(2200, 1)的测试标签