A pattern mining approach for information filtering systems

It is a big challenge to clearly identify the boundary between positive and negative streams for information filtering systems. Several attempts have used negative feedback to solve this challenge; however, there are two issues for using negative relevance feedback to improve the effectiveness of information filtering. The first one is how to select constructive negative samples in order to reduce the space of negative documents. The second issue is how to decide noisy extracted features that should be updated based on the selected negative samples. This paper proposes a pattern mining based approach to select some offenders from the negative documents, where an offender can be used to reduce the side effects of noisy features. It also classifies extracted features (i.e., terms) into three categories: positive specific terms, general terms, and negative specific terms. In this way, multiple revising strategies can be used to update extracted features. An iterative learning algorithm is also proposed to implement this approach on the RCV1 data collection, and substantial experiments show that the proposed approach achieves encouraging performance and the performance is also consistent for adaptive filtering as well.

The effect of endotracheal suction on regional tidal ventilation and end-expiratory lung volume

Purpose: To examine the impact of different endotracheal tube (ETT) suction techniques on regional end-expiratory lung volume (EELV) and tidal volume (VT) in an animal model of surfactant-deficient lung injury. Methods: Six 2-week old piglets were intubated (4.0 mm ETT), muscle-relaxed and ventilated, and lung injury was induced with repeated saline lavage. In each animal, open suction (OS) and two methods of closed suction (CS) were performed in random order using both 5 and 8 French gauge (FG) catheters. The pre-suction volume state of the lung was standardised on the inflation limb of the pressure-volume relationship. Regional EELV and VT expressed as a proportion of the impedance change at vital capacity (%ZVCroi) within the anterior and posterior halves of the chest were measured during and for 60 s after suction using electrical impedance tomography. Results: During suction, 5 FG CS resulted in preservation of EELV in the anterior (nondependent) and posterior(dependent) lung compared to the other permutations, but these only reached significance in the anterior regions (p\0.001 repeated-measures ANOVA). VT within the anterior, but not posterior lung was significantly greater during 5FG CS compared to 8 FG CS; the mean difference was 15.1 [95% CI 5.1, 25.1]%ZVCroi. Neither catheter size nor suction technique influenced post-suction regional EELV or VT compared to pre-suction values (repeated-measures ANOVA). Conclusions: ETT suction causes transient loss of EELV and VT throughout the lung. Catheter size exerts a greater influence than suction method, with CS only protecting against derecruitment when a small catheter is used, especially in the non-dependent lung.

How can soil monitoring networks be used to improve predictions of organic carbon pool dynamics and CO2 fluxes in agricultural soils?

Abstract As regional and continental carbon balances of terrestrial ecosystems become available, it becomes clear that the soils are the largest source of uncertainty. Repeated inventories of soil organic carbon (SOC) organized in soil monitoring networks (SMN) are being implemented in a number of countries. This paper reviews the concepts and design of SMNs in ten countries, and discusses the contribution of such networks to reducing the uncertainty of soil carbon balances. Some SMNs are designed to estimate country-specific land use or management effects on SOC stocks, while others collect soil carbon and ancillary data to provide a nationally consistent assessment of soil carbon condition across the major land-use/soil type combinations. The former use a single sampling campaign of paired sites, while for the latter both systematic (usually grid based) and stratified repeated sampling campaigns (5–10 years interval) are used with densities of one site per 10–1,040 km². For paired sites, multiple samples at each site are taken in order to allow statistical analysis, while for the single sites, composite samples are taken. In both cases, fixed depth increments together with samples for bulk density and stone content are recommended. Samples should be archived to allow for re-measurement purposes using updated techniques. Information on land management, and where possible, land use history should be systematically recorded for each site. A case study of the agricultural frontier in Brazil is presented in which land use effect factors are calculated in order to quantify the CO2 fluxes from national land use/management conversion matrices. Process-based SOC models can be run for the individual points of the SMN, provided detailed land management records are available. These studies are still rare, as most SMNs have been implemented recently or are in progress. Examples from the USA and Belgium show that uncertainties in SOC change range from 1.6–6.5 Mg C ha−1 for the prediction of SOC stock changes on individual sites to 11.72 Mg C ha−1 or 34% of the median SOC change for soil/land use/climate units. For national SOC monitoring, stratified sampling sites appears to be the most straightforward attribution of SOC values to units with similar soil/land use/climate conditions (i.e. a spatially implicit upscaling approach). Keywords Soil monitoring networks - Soil organic carbon - Modeling - Sampling design

Fault diagnosis of induction motor based on decision trees and adaptive neuro-fuzzy inference

This paper presents a fault diagnosis method based on adaptive neuro-fuzzy inference system (ANFIS) in combination with decision trees. Classification and regression tree (CART) which is one of the decision tree methods is used as a feature selection procedure to select pertinent features from data set. The crisp rules obtained from the decision tree are then converted to fuzzy if-then rules that are employed to identify the structure of ANFIS classifier. The hybrid of back-propagation and least squares algorithm are utilized to tune the parameters of the membership functions. In order to evaluate the proposed algorithm, the data sets obtained from vibration signals and current signals of the induction motors are used. The results indicate that the CART–ANFIS model has potential for fault diagnosis of induction motors.