No abstract

In this paper, we propose a novel hybrid text classification model based on deep belief network and softmax regression. To solve the sparse high-dimensional matrix computation problem of texts data, a deep belief network is introduced. After the feature extraction with DBN, softmax regression is employed to classify the text in the learned feature space. In pre-training procedures, the deep belief network and softmax regression are first trained, respectively. Then, in the fine-tuning stage, they are transformed into a coherent whole and the system parameters are optimized with Limited-memory BroydenFletcher-Goldfarb-Shanno algorithm. The experimental results on Reuters-21,578 and 20-Newsgroup corpus show that the proposed model can converge at fine-tuning stage and perform significantly better than the classical algorithms, such as SVM and KNN.

No abstract

With the continuing increase in the impact of human activities on ecosystems, ecologists are increasingly becoming interested in understanding the effects of nitrogen deposition on litter decomposition. At present, numerous studies have investigated the effects of single form of nitrogen fertilization on litter decomposition in forest ecosystems. However, forms of N deposition vary, and changes in the relative importance of different forms of N deposition are expected in the future. Thus, identifying the effects of different forms of N deposition on litter decomposition in forest ecosystems is a pressing task. In this study, two dominant litter types were chosen from Zijin Mountain in China: Quercus acutissima leaves from a late succession broad-leaved forest and Pinus massoniana needles from an early succession coniferous forest. The litter samples were incubated in microcosms with original forest soil and treated with four different forms of nitrogen fertilization [NH 4 + , NO 3 À , CO(NH 2 ) 2 , and a mix of all three]. During a 5-month incubation period, litter mass losses, soil pH values, and soil enzyme activities were determined. Results show that all four forms of nitrogen fertilization significantly accelerate litter decomposition rates in the broadleaf forest, while only two forms of nitrogen fertilization [i.e., mixed nitrogen and CO(NH 2 ) 2 ] significantly accelerate litter decomposition rates in the coniferous forest. Litter decomposition rates with the mixed nitrogen fertilization were higher than those in any single form of nitrogen fertilization. All forms of nitrogen fertilization enhanced soil enzyme activities (i.e., catalase, cellulase, invertase, polyphenol oxidase, nitrate reductase, urease, and acid phosphatase) during the litter decomposition process for the two forest types. Soil enzyme activities under the mixed nitrogen fertilization were higher than those under any single form of nitrogen fertilization. These results suggest that the type and activity of the major degradative enzymes involved in litter decomposition vary in different forest types under different forms of nitrogen fertilization. They also indicate that a long-term consequence of N depositioninduced acceleration of litter decomposition rates in subtropical forests may be the release of carbon stored belowground to the atmosphere.

Motivation Deep neural network (DNN) algorithms were utilized in predicting various biomedical phenotypes recently, and demonstrated very good prediction performances without selecting features. This study proposed a hypothesis that the DNN models may be further improved by feature selection algorithms. Results A comprehensive comparative study was carried out by evaluating 11 feature selection algorithms on three conventional DNN algorithms, i.e. convolution neural network (CNN), deep belief network (DBN) and recurrent neural network (RNN), and three recent DNNs, i.e. MobilenetV2, ShufflenetV2 and Squeezenet. Five binary classification methylomic datasets were chosen to calculate the prediction performances of CNN/DBN/RNN models using feature selected by the 11 feature selection algorithms. Seventeen binary classification transcriptome and two multi-class transcriptome datasets were also utilized to evaluate how the hypothesis may generalize to different data types. The experimental data supported our hypothesis that feature selection algorithms may improve DNN models, and the DBN models using features selected by SVM-RFE usually achieved the best prediction accuracies on the five methylomic datasets. Availability and implementation All the algorithms were implemented and tested under the programming environment Python version 3.6.6. Supplementary information Supplementary data are available at Bioinformatics online.