Innovative propensity with a fuzzy multicriteria approach

Purpose Although R&D plays a crucial role in innovativeness and R&D expenditures is the most widely used tool to measure the level of innovativeness of companies, other variables and inputs may be equally interesting. The purpose of this paper is to define an innovative propensity index (IPI) which considers these variables and allows the identification of those companies which have a higher propensity to implement different types of innovativeness. Design/methodology/approach Taking into account, the different criteria that may be considered in an IPI and that the perception of the relative importance of each criterion is subjective, the use of an innovativeness multicriteria decision methodology has been considered appropriate. In particular, an IPI is built from the weighting of the criteria through FAHP methodology. Data mining techniques are subsequently used to establish a non-supervised ranking (clustering) of a sample of firms, considering their IPI values. Findings The application of an IPI to a sample of 1,639 companies operating in different industrial sectors has helped us to find out that this index is useful for identifying those companies which really show an increased innovative capacity. A comparative analysis by sectors has shown that although there are companies from all sectors with a high innovative propensity, the proportion increases in more technological sectors. Moreover, it has been observed that in companies with higher net personnel expenses and high productivity level the innovative propensity is also higher. Originality/value The criteria used to build the index affects innovativeness individually, but the value of the analysis lies in its multicriteria approach and use of fuzzy logic. The validation of the index in a wide sample of firms is another outstanding aspect of the analysis.

To Detect Irregular Trade Behaviors In Stock Market By Using Graph Based Ranking Methods

: To detect the irregular trade behaviors in the stock market is the important problem in machine learning field. These irregular trade behaviors are obviously illegal. To detect these irregular trade behaviors in the stock market, data scientists normally employ the supervised learning techniques. In this paper, we employ the three graph Laplacian based semi-supervised ranking methods to solve the irregular trade behavior detection problem. Experimental results show that that the un-normalized and symmetric normalized graph Laplacian based semisupervised ranking methods outperform the random walk Laplacian based semi-supervised ranking method.

Concept Grounding to Multiple Knowledge Bases via Indirect Supervision

We consider the problem of disambiguating concept mentions appearing in documents and grounding them in multiple knowledge bases, where each knowledge base addresses some aspects of the domain. This problem poses a few additional challenges beyond those addressed in the popular Wikification problem. Key among them is that most knowledge bases do not contain the rich textual and structural information Wikipedia does; consequently, the main supervision signal used to train Wikification rankers does not exist anymore. In this work we develop an algorithmic approach that, by carefully examining the relations between various related knowledge bases, generates an indirect supervision signal it uses to train a ranking model that accurately chooses knowledge base entries for a given mention; moreover, it also induces prior knowledge that can be used to support a global coherent mapping of all the concepts in a given document to the knowledge bases. Using the biomedical domain as our application, we show that our indirectly supervised ranking model outperforms other unsupervised baselines and that the quality of this indirect supervision scheme is very close to a supervised model. We also show that considering multiple knowledge bases together has an advantage over grounding concepts to each knowledge base individually.

A Structural Equation Modeling Based Approach for Identifying Key Descriptors in Microstructural Materials Design

In design of advanced heterogeneous materials system, microstructures play an important role as a link between processing and material properties. An accurate and efficient representation of material microstructures is necessary. Our prior work applied a supervised ranking algorithm to identify key microstructure descriptors, however the approach falls short in identifying redundancy in descriptors and is not reliable when the training sample size is small. In this paper, we propose a Structural Equation Modeling (SEM) based approach to identify significant microstructure descriptors based on either correlation functions (CF) or material properties, or both. By building a reflective structural model, we are able to deal with high correlations among all candidate descriptors, gain more insights into their relations, and identify latent factors for categorizing microstructure features. The proposed approach begins with an Exploratory Factor Analysis (EFA) for grouping and reducing descriptors to determine the proper structure of microstructure descriptors as indicators of latent factors. The SEM analysis is then applied to identify the key descriptors using the Partial Least Squares (PLS) algorithm. The nanodielectric system with epoxy-nanosilica is used as an example to illustrate and validate the proposed approach. The potential use of identified key microstructure descriptors for optimal design of microstructural materials is discussed.