Pages 303–308
Abstract
Research summary supplied by the authors
Motivation: A fast-biclustering algorithm for binary datasets is presented here. In order to significantly reduce the computational overhead in comparison to the state-of-the-art BiBit algorithm that uses repeated bit-wise AND operations followed by comparisons, we have introduced a novel encoding and search strategy in our method. We can use biclustering in the leather industry for production processes optimization, productivity enhancement, and improve supply chain management. This can be done by grouping similar inputs (like raw hides, chemicals) and conditions (for example process times, temperatures). Inspite of wide use of biclustering in bioinformatics, its application in the leather industry includes adapting the technique to manufacturing and supply chain data. One of its applications is in the leather industry is optimizing the tanning process. In order to achieve desired leather properties in the leather manufacturing technique, it uses different types of chemicals and specific conditions like temperature, duration and pH levels. This can be done if we keep different batches of raw hides in rows and several combinations of chemical agents and process parameter in columns. This data matrix can be used to find biclusters which are subset of hides and subsets of process or chemical conditions. In this way finding biclusters will help to find high quality soft and durable leather which in turns also minimize waste.