The performance of BLMSumm: Distinct languages with antagonistic domains and varied compressions

The article describes the BLMSumm summarizer, whose aim is to use various local search methods and Metaheuristics to create summaries. Summarization is treated as an issue of optimization and an attribute identification method based on bipartite graphs is presented. BLMSumm is enhanced with a heuristic for attribute selection, the purpose of which is to make it language independent and then compare it to both professional algorithms and those in the literature. The results obtained from the experiments and evaluated using the Rouge tool are promising.