Ensemble classification based on generalized additive models

Generalized additive models (GAMs) are a generalization of generalized linear models (GLMs) and constitute a powerful technique which has successfully proven its ability to capture nonlinear relationships between explanatory variables and a response variable in many domains. In this paper, GAMs are proposed as base classifiers for ensemble learning. Three alternative ensemble strategies for binary classification using GAMs as base classifiers are proposed: (i) GAMbag based on Bagging, (ii) GAMrsm based on the Random Subspace Method (RSM), and (iii) GAMens as a combination of both. In an experimental validation performed on 12 data sets from the UCI repository, the proposed algorithms are benchmarked to a single GAM and to decision tree based ensemble classifiers (i.e. RSM, Bagging, Random Forest, and the recently proposed Rotation Forest). From the results a number of conclusions can be drawn. Firstly, the use of an ensemble of GAMs instead of a single GAM always leads to improved prediction performance. Secondly, GAMrsm and GAMens perform comparably, while both versions outperform GAMbag. Finally, the value of using GAMs as base classifiers in an ensemble instead of standard decision trees is demonstrated. GAMbag demonstrates performance comparable to ordinary Bagging. Moreover, GAMrsm and GAMens outperform RSM and Bagging, while these two GAM ensemble variations perform comparably to Random Forest and Rotation Forest. Sensitivity analyses are included for the number of member classifiers in the ensemble, the number of variables included in a random feature subspace and the number of degrees of freedom for GAM spline estimation.     

Weak signal identification with semantic web mining

We investigate an automated identification of weak signals according to Ansoff to improve strategic planning and technological forecasting. Literature shows that weak signals can be found in the organization’s environment and that they appear in different contexts. We use internet information to represent organization’s environment and we select these websites that are related to a given hypothesis. In contrast to related research, a methodology is provided that uses latent semantic indexing (LSI) for the identification of weak signals. This improves existing knowledge based approaches because LSI considers the aspects of meaning and thus, it is able to identify similar textual patterns in different contexts. A new weak signal maximization approach is introduced that replaces the commonly used prediction modeling approach in LSI. It enables to calculate the largest number of relevant weak signals represented by singular value decomposition (SVD) dimensions. A case study identifies and analyses weak signals to predict trends in the field of on-site medical oxygen production. This supports the planning of research and development (R&D) for a medical oxygen supplier. As a result, it is shown that the proposed methodology enables organizations to identify weak signals from the internet for a given hypothesis. This helps strategic planners to react ahead of time.     

Interactions between simultaneously activated behavioral systems in the rat.

Interactions between electrically induced attack and teeth-chattering from 1 electrode and grooming from another were examined in male albino rats. The interaction between electrically induced attacks and deprivation-induced feeding, as well as the effect of food deprivation on attack, was also studied. Results indicate that attack appears to be a dominant response, for it suppressed grooming and feeding at a low level of activation. On the other hand, it was not affected by simultaneously induced grooming or feeding. However, food deprivation decreased the threshold for attack, leaving attack latency, attack form, or bite targets unaffected. Teeth-chattering, suggested to be related to attack and flight, was also a dominant response. Results suggest that interactions between behavioral systems are in favor of the systems that must act acutely on activation in order to survive. Apparently, the regulations governing these interactions are represented in the functional organization of the brain. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

In vitro gas production profiles and fermentation end‐products in processed peas,lupins and faba beans

BACKGROUND: An experiment was carried out to establish whether using a pre‐compacting device (expander) changes the contribution of dry matter (DM) and degradative behaviour of peas, lupins and faba beans over the different fractions (non‐washable fraction, NWF; insoluble washable fraction, ISWF; soluble washable fraction, SWF). Samples of the entire concentrate ingredients (WHO ingredients) and their different fractions (NWF, ISWF and SWF) were subjected to three processes (Retsch milling, R; expander treatment, E; expander‐pelleting, EP) and their fermentation characteristics were evaluated using an in vitro gas production technique. RESULTS: In peas and faba beans, both the E and EP processes increased the size of the NWF (P < 0.05) and decreased the size of the SWF compared with the R process. The maximum fractional rate of gas production in the first phase of fermentation was higher in the E and EP samples than in the R samples (P < 0.05). In lupins and faba beans the E and EP processes shifted the pattern of fermentation towards a more glucogenic fermentation, as represented by a lower non‐glucogenic to glucogenic ratio (NGR). Ammonia production (NH₃‐N) in the E and EP samples was significantly (P < 0.05) lower than that in the R samples. CONCLUSION: It is concluded that the E and EP processes provide a certain level of protection against ruminal breakdown to dietary protein and shift the pattern of fermentation towards a more glucogenic fermentation. Copyright © 2008 Society of Chemical Industry     

Untersuchungen zur Verteilung des Klebstoffes im Bereich der Leimfuge mittels Neutronenradiographie und Mikroskopie

The permeation behaviour of different types of adhesives (urea-formaldehyde resin, dispersion glue (PVA) and Polyurethane) was investigated by light microscopy, scanning electron microscopy and neutron radiography. Clear differences in the permeation behaviour of the adhesives could be determined. Urea resin showed under the chosen conditions the strongest permeation of wood. The glue allocation was quantified and the permeation visualised by means of neutron radiography. By this a differentiation between wood and the glue joint was possible. The penetration depth of the glue was conspicuously different between spring wood and late wood areas.     

Lifetime Assessment of Load‐Bearing Polymer Glasses: An Analytical Framework for Ductile Failure

The most widespread application of polymers in structural applications is their use as pipe material for e.g., gas distribution systems. Pipes have a design lifetime of typically 50 years, which rules out real‐time lifetime assessment methods. Here, an engineering approach is presented, which makes it possible to predict long‐term ductile failure of loaded glassy polymers based on short‐term tests. The approach is based upon the hypothesis that failure is governed by accumulation of plastic deformation up to a critical strain. A pressure‐modified Eyring relation is employed to calculate the accumulation of plastic strain for any simple loading geometry. It is demonstrated that the approach can produce accurate quantitative time‐to‐failure predictions for loaded PC specimens and uPVC pipe segments.

     

Analyzing existing customers’ websites to improve the customer acquisition process as well as the profitability prediction in B-to-B marketing

We investigate the issue of predicting new customers as profitable based on information about existing customers in a business-to-business environment. In particular, we show how latent semantic concepts from textual information of existing customers’ websites can be used to uncover characteristics of websites of companies that will turn into profitable customers. Hence, the use of predictive analytics will help to identify new potential acquisition targets. Additionally, we show that a regression model based on these concepts is successful in the profitability prediction of new customers. In a case study, the acquisition process of a mail-order company is supported by creating a prioritized list of new customers generated by this approach. It is shown that the density of profitable customers in this list outperforms the density of profitable customers in traditional generated address lists (e.g. from list brokers). From a managerial point of view, this approach supports the identification of new business customers and helps to estimate the future profitability of these customers in a company. Consequently, the customer acquisition process can be targeted more effectively and efficiently. This leads to a competitive advantage for B2B companies and improves the acquisition process that is time- and cost-consuming with traditionally low conversion rates.     

Tracing enteric viruses in the European berry fruit supply chain

In recent years, numerous foodborne outbreaks due to consumption of berry fruit contaminated by human enteric viruses have been reported. This European multinational study investigated possible contamination routes by monitoring the entire food chain for a panel of human and animal enteric viruses.     

Visual word recognition in bilinguals: Evidence from masked phonological priming.

Bilingual written language representation was investigated with the masked phonological priming paradigm. Pseudohomophonic and control primes of French target words were used to show that Dutch–French bilinguals exhibit the same pattern of phonological and orthographic priming as native French speakers, which suggests that the same processes underlie first- and second-language processing. It was also found that for bilinguals, but not monolinguals, it is possible to prime a target word of the second language with a homophonic stimulus (either word or nonword) of the first language. This interlingual phonological priming effect was of the same size as the intralingual priming effect. Implications for theories of bilingual written language representation and for the interpretation of the masked phonological priming paradigm are discussed. (PsycINFO Database Record (c) 2010 APA, all rights reserved)