Inductive Model Generation for Text Classification Using a Bipartite Heterogeneous Network

Algorithms for numeric data classification have been applied for text classification. Usually the vector space model is used to represent text collections. The characteristics of this representation such as sparsity and high dimensionality sometimes impair the quality of general-purpose classifiers. Networks can be used to represent text collections, avoiding the high sparsity and allowing to model relationships among different objects that compose a text collection. Such network- based representations can improve the quality of the classification results. One of the simplest ways to represent textual collections by a network is through a bipartite heterogeneous network, which is composed of objects that represent the documents connected to objects that represent the terms. Heterogeneous bipartite networks do not require computation of similarities or relations among the objects and can be used to model any type of text collection. Due to the advantages of representing text collections through bipartite heterogeneous networks, in this article we present a text classifier which builds a classification model using the structure of a bipartite heterogeneous network. Such an algorithm, referred to as IMBHN （Inductive Model Based on Bipartite Heterogeneous Network）, induces a classification model assigning weights to objects that represent the terms for each class of the text collection. An empirical evaluation using a large amount of text collections from different domains shows that the proposed IMBHN algorithm produces significantly better results than k-NN, C4.5, SVM, and Naive Bayes algorithms.     

Flow map layout via spiral trees

Flow maps are thematic maps that visualize the movement of objects, such as people or goods, between geographic regions. One or more sources are connected to several targets by lines whose thickness corresponds to the amount of flow between a source and a target. Good flow maps reduce visual clutter by merging (bundling) lines smoothly and by avoiding self-intersections. Most flow maps are still drawn by hand and only few automated methods exist. Some of the known algorithms do not support edge-bundling and those that do, cannot guarantee crossing-free flows. We present a new algorithmic method that uses edge-bundling and computes crossing-free flows of high visual quality. Our method is based on so-called spiral trees, a novel type of Steiner tree which uses logarithmic spirals. Spiral trees naturally induce a clustering on the targets and smoothly bundle lines. Our flows can also avoid obstacles, such as map features, region outlines, or even the targets. We demonstrate our approach with extensive experiments.     

Impact of oxygen concentration during the deposition of window layers on lowering the metastability effects in Cu(In,Ga)Se2/CBD Zn(S,O) based solar cell

The purpose of the present paper is to focus on the impact of oxygen gas partial pressure during the sputtering of i‐ZnO and ZnMgO on the transient behavior of Cu(In,Ga)Se₂ (CIGSe) based solar cells parameters when a CBD‐Zn(S,O) buffer layer is used. Based on electrical characterization of cells, it is observed that the effect of light soaking is different on J–V characteristics depending on whether oxygen is or is not present during the first deposition time of the i‐ZnO or ZnMgO layers. In fact, when cells are prepared with standard i‐ZnO, the efficiencies are very low and a pronounced transient behavior is observed. However, when the first 10 nm of i‐ZnO or ZnMgO is formed by sputtered layer without adding oxygen during the process, depending on the thickness of the buffer layer, the transient effects strongly decreases. It is then possible to get stable cells reaching efficiencies quite similar to the CdS reference cells, especially with ZnMgO, without any post‐treatments. Copyright © 2015 John Wiley & Sons, Ltd.     

Effect of dehydroxylation on the specific heat of simple clay mixtures

Specific heats of four clays (standard reference kaolins, commercial kaolin and montmorillonite) before and after dehydroxylation have been measured. The results were compared with handbook data for the thermal chemical properties of solids. Good agreement has been obtained for the reference kaolin before any thermal treatment. Then, following thermal treatments at 500 °C, 600 °C and 700 °C, dehydroxylation leads to a progressive decrease of heat capacity per unit mass. After dehydroxylation, heat capacity values for all the studied materials are rather similar and agree closely with those estimated by the rule of mixtures. Finally, an empirical relation describing the specific heat capacity (C) in J kg⁻¹ K⁻¹ of dehydroxylated kaolin from 40 °C to 1100 °C is proposed: C = 1128 + 0.102T − 36 × 10⁶T⁻² where T is in K.     

The structure of thermomigration channels in silicon

We have studied deep and through channels formed in n-Si wafers by means of thermal migration of thin discrete aluminum zones. The region of thermal migration channels was investigated using selective chemical etching of silicon in combination with secondary ion mass spectrometry for the analysis of impurity distributions. It is established that the channels are surrounded by two distinct peripheral shells.     

Certifying circuits in Type Theory

We investigate how to take advantage of the particular features of the calculus of inductive constructions in the framework of hardware verification. First, we emphasize in a short case study the use of dependent types and of the constructive aspect of the logic for specifying and synthesizing combinatorial circuits. Then, co-inductive types are introduced to model the temporal aspects of sequential synchronous devices. Moore and Mealy automata are co-inductively axiomatized and are used to represent uniformly both the structures and the behaviors of the circuits. This leads to clear, general and elegant proof processes as is illustrated on the example of a realistic circuit: the ATM Switch Fabric. All the proofs are carried out using Coq.Accepted in revised form 29 February 2004 by C.B. Jones     

Fungal fucoidanase production by solid-state fermentation in a rotating drum bioreactor using algal biomass as substrate

Fucoidanase enzymes able to degrade fucoidan were produced by solid-state fermentation (SSF). The fermentation assays were initially carried out in a laboratory-scale rotating drum bioreactor, and two fungal strains (Aspergillus niger PSH and Mucor sp. 3P) and three algal substrates (untreated, autohydrolyzed, and microwave processed seaweed Fucus vesiculosus) were evaluated. Additionally, fermentations were carried out under rotational (10 rpm) and static conditions in order to determine the effect of the agitation on the enzyme production. Agitated experiments showed advantages in the induction of the enzyme when compared to the static ones. The conditions that promoted the maximum fucoidanase activity (3.82 U L^?1) consisted in using Mucor sp. 3P as fungal strain, autohydrolyzed alga as substrate, and the rotational system. Such conditions were subsequently used in a 10 times larger scale rotating drum bioreactor. In this step, the effect of controlling the substrate moisture during the enzyme production by SSF was investigated. Moreover, assays combining the algal substrate with an inert support (synthetic fiber) were also carried out. Fermentation of the autohydrolyzed alga with the moisture content maintained at 80% during the fermentation with Mucor sp. 3P gave the highest enzyme activity (9.62 U L^?1).     

New statistical software for the proportional hazards model with current status data

Currently, there are some statistical software packages such as intcox, survBayes, and BITE that are designed to analyze current status data (in which failure times are known to be either left- or right-censored) based on the proportional hazards model. They, however, either fail to directly provide standard errors for the estimated regression parameters or require frailty terms. As a result, practitioners often analyze their data using packages for right-censored data. By so doing, they mistreat left-censored observations as exact. This paper describes our newly developed statistical software for the proportional hazards model with current status data. The software is implemented in the R and C languages and consists of the following two simple steps: (a) find MLEs of the regression parameter and the cumulative hazard function; (b) compute the variance-covariance matrix of the regression parameter estimator by using the generalized missing information principle (GMIP) developed by Kim [Kim, J.S., 2003b. Efficient estimation for the proportional hazards model with left-truncated and Case 1 interval-censored data. Statista Sinica 13 (2), 519-537]. Our simulation study results show that our method works well in terms of bias, standard error, and power. By treating current status data as right-censored data, we also show the discrepancy in terms of bias, standard error, and power. Real examples are provided to illustrate the use of the software. This method can be extended to both general interval-censored data and truncated and interval-censored data.