web mining based extraction of problem solution ideas

A new idea web mining approach that automatically identifies new technological ideas extracted from the internet. Existing approaches for identifying ideas from texts are adapted to the use of the internet. An automated assignment of ideas to different domains is realized that enables the extraction of ideas from the full text of internet documents by use of an optimized set of parameter values. For future work, the used knowledge based classification approach should be replaced with a semantic classification approach. Well-known methods e.g. latent semantic indexing or non-negative matrix factorization probably are suited to identify new ideas with a better performance than the proposed approach because they consider aspects of meaning rather than aspects of words. A further avenue of research focuses on the automated translation of extracted text patterns from websites.