HADOOP DATSETS BINARY CLASSIFICATION

The Hadoop dataset consists of multiple folders, where each folder corresponds to a unique application ID and contains the raw log files generated by that specific application. An additional file, abnormal_label, is provided to indicate which log sequences correspond to abnormal executions, allowing the assignment of anomaly labels to the records. Two versions of the dataset were prepared: the first version follows the original structure and labeling provided in the official Hadoop logs released on the LogPAI repository, while the second version incorporates the corrected labels proposed in the public discussion on the LogDelta issue tracker. This dual-version design preserves both the original dataset integrity and the improved labeling quality suggested by the community, ensuring greater reliability for downstream anomaly detection tasks.1- Original row datasets: https://zenodo.org/records/8196385 2- 1st verion based on https://github.com/logpai/loghub/tree/master/Hadoop >> named HADOOP_V1.csv 3- 2d verion fix incorrect label based on https://github.com/logpai/loghub/issues/56 >> named HADOOP_V2.csvTHESE DATA SET LABELED AS BINARY ( Normal , Anomaly).