LearnAndReduce MWIS Trainingdata

Learn and Reduce Dataset This dataset consists of vertex-labeled graphs used to train machine learning models for data reduction screening. This is used as a preprocessing step to solve the Maximum Weight Independent Set problem. The original folder contains all instances and reduction data using the computationally cheap reduction rules. The graphs and reduction data are stored as two separate CSV files, with the suffix _original_graph.csv and _original_reduction_data.csv. Similarly, the kernel folder contains the non-empty reduced instances. Here, we use the suffix _kernel_graph.csv and _kernel_reduction_data.csv. The graph files have two columns, source and target. Each undirected edge {u,v} in the graph appears as both u;v and v;u.