UPMVM datasets

UPMVM used three datasets named UD1, UD2 and UD3. UD1 is primarily used to collect and retrieve 280 poetry meters (rhythmic patterns [بحر]) and their corresponding feet. Other uses of this dataset include the design of DFA state function sequences with terminal state information to align the identified verse meters. UD2 is collected from [GitHub - sayedzeeshan/Aruuz] and updated. This update process involves the parsing and tokenization of the UD2 dataset. Arud rules are then applied to each token for its discretization, resulting in a tok-taqti which is then added to the vocabulary if it is not already present. This dataset UD2 contains 80,000 Urdu words and their diacritics, language information, and word variations. The third dataset UD3 is used to evaluate the proposed model. UD3 contains 500 verses including structure-based genre, collected from published books.

UPMVM采用了三个数据集，分别命名为UD1、UD2和UD3。UD1主要用于收集和检索280种诗律（韵律模式[ بحر]）及其对应的脚步。此外，此数据集还用于设计包含终止态信息的DFA状态函数序列，以对识别出的诗律进行对齐。UD2数据集来源于[GitHub - sayedzeeshan/Aruuz]，并进行了更新。该更新过程包括对UD2数据集的解析和分词。随后，对每个标记应用Arud规则进行离散化，形成tok-taqti，若其尚未存在于词汇表中，则将其添加。UD2数据集包含80,000个乌尔都语单词及其变音符号、语言信息和单词变体。第三个数据集UD3用于评估所提出的模型。UD3包含500首诗句，包括基于结构的体裁，这些诗句均选自已出版的书籍。