Uniformly subsampled ensemble (USE) for churn management: Theory and implementation

The present paper explores the possible application of a new ensemble model. The model, which is based on multiple SVM classifiers, is employed to address churner identification problems in the mobile telecommunication industry, a sector in which the role of customer retention program becomes increasingly important due to its very competitive business environment. In particular, the current study introduces a uniformly subsampled ensemble (USE) model of SVM classifiers, not only to reduce the computational complexity of large-scale data, but also to boost the reliability and accuracy of calibrated models on data sets with highly skewed class distributions. According to our experiments, the performance of the USE SVM model is superior compared to all single and ensemble models. It is more scalable than well-known ensemble models as well.     

Measuring the success of retention management models built on churn probability,retention probability,and expected yearly revenues

We claim that often marketers have not all the information to develop various marketing campaign models. For example, marketers may have sufficient information to build a model for predicting possible churners, while they may have no clues of which customers are most likely to accept a retention campaign. In this paper, we first show that the information useful for a successful churner prediction model alone is not sufficient to develop a successful retention marketing program. In such a case, we claim that only theory-based simulation approach is feasible. In particular, it is claimed that optimal retention management models should consider not only churn probability but also retention probability and expected revenues from target customers. To validate our claim, we develop and compare five retention management models based on churn probability, retention probability, expected revenues, and combination of these models along with different evaluation metrics. Our experimental results show that the retention management model with the highest accuracy in predicting possible churners is not necessarily optimal because it does not consider the probability of accepting retention promotions. In contrast, the retention management model based on both churn and retention probability is the best in terms of predicting customers who are most likely to positively respond to retention promotions. Ultimately, the model based on expected yearly revenue of customers accrues the highest revenues across most target points, making it the best model out of five churn management models.     

Science in Brazil. Part 1: A macro-level comparative study

Summary In the present paper, the evolution of publication activity and citation impact in Brazil is studied for the period 1991-2003. Besides the analysis of trends in publication and citation patterns and of national publication profiles, an attempt is made to find statistical evidences of the relation between international co-authorship and both research profile and citation impact in the Latin American region. Despite similarities and strong co-publication links with the other countries in the region, Brazil has nonetheless a specific research profile, and forms the largest potential in the region.     

Patent evolution in relation to public/private R&D investment and corporate profitability: Evidence from the United States

Summary In this paper we analyze the (historical) co-evolution of technological development and economic progress (by relating public and private R&D investment, patenting, and corporate profitability). We relate to the work ofSchmookler(1966),Griliches(1990),Pakes&Griliches(1980) andPakes(1986) who all have studied the techno-economic interplay by considering patents as in indicator of technological performance. We use United States industry and government data over the period 1953-1998 (45 years). Co-evolution analysis over this period reveals a strong interdependency among the variables. Patent evolution is strongly related to the development of private R&D and corporate profitability; the levels of public and private R&D expenditure in combination with the level of technological output (i.e. patents) have a strong predictive and explanatory power towards corporate profitability (R² value of 94.9%). Causality tests reveal a joint determination between R&D investment and corporate profitability (L=2; p<0.01).     

Assembly of ordered ZnO porous thin films by cooperative assembly method using polystyrene spheres and ultrafine ZnO particles

Ordered ZnO porous thin films were fabricated by cooperative assembly method using polystyrene sphere (PS) and ultrafine ZnO particles, in which ultrafine ZnO particles were directly assembled in the voids of PS while the template was being assembled by capillary forces. The influence of experimental parameters, such as evaporation temperature, ZnO concentration and the concentration ratio of PS/ZnO on morphology of the porous structure was mainly studied. The results showed that an ordered porous structure could be obtained by this method. X-ray diffraction (XRD) spectra indicated the porous ZnO thin film was wurtzite structure. The transmissivity decreased with the decrease of wavelength, but still kept above 80% beyond the wavelength of 550 nm. Optical band gap of the ZnO thin film was 3.13 eV.     

Electrochemical performance of Y2O3/NiO cathode in the molten Li0.62/K0.38 carbonates eutectics

The preparation and subsequent oxidation of Ni cathodes modified by impregnation with yttria were evaluated by surface and bulk analysis. The electrochemical behavior of Y₂O₃/NiO cathodes was also evaluated in a molten 62 mol% Li₂CO₃ + 38 mol% K₂CO₃ eutectic at 650 °C by electrochemical impedance spectroscopy (EIS) as a function of yttria content and immersion time under the standard cathode gas condition (CO₂:O₂ = 67:33%). The stability tests of Y₂O₃/NiO cathodes showed that the yttria additive could dramatically reduce the solubility of NiO in the eutectic molten Li/K carbonates due to the preferential dissolution of yttria. The loss of yttria was confirmed by chemical analysis and X-ray diffraction (XRD). The Y₂O₃/NiO cathodes showed higher catalytic activity for oxygen reduction and lower dissolution of NiO than the pure NiO cathode. The cathode material with 1.0 wt.% of yttria showed the optimum behavior.     

A neural network for predicting the stability of RNA/DNA hybrid duplexes

A back propagation (BP) neural network (NN) method has been developed to predict the stability of RNA/DNA duplex. Calculated T_m of hybrid formation with the present parameters fits the experimental values within reasonable errors (AD=1.2887 K, SEP=1.782, R²=0.9673 for NN1; AD=2.1388 K, SEP=3.506, R²=0.8889 for NN2). And it has the advantage that the determinations of thermodynamic parameters (TP) are not needed.     

Cross-sectional nanoprobing fault isolation technique on submicron devices

With continuous scaling on CMOS device dimensions, it is becoming increasingly challenging for conventional failure analysis (FA) methods to identify the failure mechanism at the circuit level in an integrated chip. Scanning Electron Microscopy (SEM) based nanoprobing is becoming an increasingly critical tool for identifying non-visual failures via electrical characterization in current electrical FA metrology for fault isolation since 2006 Toh et al. (2007), Shen et al. (2007), Ng et al. (2012) . Currently, most of the nanoprobing fault isolation is nanoprobe in top-down planar direction, such as nanoprobe on via, contact and metal line. This paper focused on fault isolation of sub-micron devices by nanoprobing on a cross-sectional plane. This is a new application area; it is very useful for sample that cannot perform fault isolation with conventional top-down planar nanoprobing, especially on non-volatile memory that with single transistor memory array that arrange in a vertical direction, such as Magnetic Random Access Memory (MRAM), Phase-Change Random Access Memory (PC-RAM), flash memory and etc.     

Optimal Design of Blocked and Split-Plot Experiments for Fixed Effects and Variance Component Estimation

In general, modeling data from blocked and split-plot response surface experiments requires the use of generalized least squares and the estimation of two variance components. The literature on the optimal design of blocked and split-plot response surface experiments, however, focuses entirely on the precise estimation of the fixed factor effects and completely ignores the necessity to estimate the variance components as well. To overcome this problem, we propose a new Bayesian optimal design criterion which focuses on both the variance components and the fixed effects. A novel feature of the criterion is that it incorporates prior information about the variance components through log-normal or beta prior distributions. The resulting designs allow for a more powerful statistical inference than traditional optimal designs. In our algorithm for generating optimal blocked and split-plot designs, we implement efficient quadrature approaches for the numerical approximation of the new optimal design criterion. Supplementary materials for this article are available online.     

Can openness to ICT and scientific research predict the ICT skills and ICT use of bachelor's students?

In this article, we focus on the question of why some bachelor's students have more proficiency in ICT skills and use computers and the Internet more frequently than other students do. To answer this question, we rely on the core variables of the TAM (Technology Acceptance Model), on the orientation of students toward scientific research, and on contextual factors. TAM has been applied successfully in research before, but in this article, the model is expanded with variables regarding the orientation of students toward scientific research, and is geared towards the relationship to basic ICT skills and the sustained use of these skills. In order to attain a more reliable instrument to measure proficiency in ICT skills and the frequency of ICT use, several new instruments are constructed. The results show that this model, a combination of four basic TAM variables, together with the identity commitment of students focused on the appreciation of scientific research and some context variables, can explain a substantial proportion of the variance in different ICT skills, but considerably less of the frequency of ICT use.