Using artificial neural networks for mapping of science and technology: Application to patents analysis

The article deals with a particular neural network mapping innovation, which is called MultiSOM. MultiSOM was firstly introduced for the information retrieval purposes. MultiSOM is the multi-map extension of the Kohonen SOM algorithm. This last concept technically supposes unsupervised classification and learning capabilities, and pays pretty importance to spatial order in the representation of data. The MultiSOM introduces the concepts of viewpoints and dynamics into the information analysis concept with its multi-maps displays and its inter-map communication process. The dynamic information exchange between maps can be exploited by an analyst in order to perform cooperative deduction between several different analyses that have been performed on the same data. The experimental context of the paper is constituted by a patent database of 1000 patents related to oil engineering. The patents structure and the patents fields semantics are firstly exploited in order to generate different viewpoints corresponding to different areas of interest for the analysts. In the experiment the selected viewpoints correspond to uses, advantages, patentees, and titles subfields of the patents. The indexing vocabulary of each viewpoint is automatically extracted of its related textual contents in the patents through a full text analysis. The resulting vocabulary is then used to rebuild patents descriptions regarding each viewpoint. These descriptions are finally classified through the unsupervised MultiSOM algorithm resulting in as much different maps as viewpoints. The section 1 of the article presents the Kohonen self-organizing maps (SOM) and their main applications in mapping of science and technology. Section 2 describes the basic improvements of SOM that have been set up for computerized analysis. Sections 3 deals with MultiSOM, the multi-map innovation of the SOM algorithm. It focuses on the specific features of the MultiSOM algorithm, especially on the viewpoint definition, on the map labelization strategies, the map on-line generalization, and the inter-map communication model. The context of the experiment on the oil engineering patents and the preprocessing of these latter will be described in the section 4. This section 4 also contains a short comparison of the obtained results with more K-means classification results. Section 5 describes the connection of MultiSOM with the Gallois lattice symbolic techniques and gives an empirical evaluation of its added value. The conclusions are finally exposed.