Papers by Keyword: LDA

Abstract: To the problem of Chinese micro-blogging topic tracking, a method combined LDA model and Bagging of ensemble learning was proposed. The method firstly used the LDA hidden topic modeling, effectively solved the issue that the dataset’s sparsity of the short text, then made the C4.5 decision tree as a weak classifier, through examples resampling to obtain multiple training set, compounding the training sets according to the voting rule, and ultimately getting the similarity of the micro-blogging topic. Experiments show that, compared with the model based on single vector model, classical TF-IDF and the tracking method of C.45Bagging similarity computing, this method have a better performance on precision, recall ratio and F1 value.

Abstract: In machine olfaction or electronic nose, sensor optimization is important to enhance pattern recognition efficiency and reduce redundant information. Highly correlated response of one sensor to two different odors implies less contribution of this sensor to the classification of these two odors. Variance difference is a significant index to measure the similarity of sensor responses. A sensor optimization method based on variance difference is proposed in this paper; both the average value of variance difference and cluster analysis of variance difference matrix were considered to identify several possible sensor subsets. Six Chinese herbal medicines and linear discrimination analysis (LDA) were applied to test the classification results in order to determine the best subset. LDA results indicated that the optimized sensor subset performed well in classification of the six Chinese medicines. The proposed sensor array optimization method could be applied to other kinds of odors classification as a novel method.

Abstract: Aiming at the target detection technique is time-consuming, noise and other issues, improved the application of FSVM in target detection, a piecewise fuzzy membership function is proposed based on LDA algorithm. The algorithm projected the original high-dimensional samples onto a one-dimensional space by LDA ,the original sample’s fuzzy weights is segment calculated based on the distribution of the projection point, reduce the impact of noise and outliers on classification results. In the simulation experiments, this method can effectively reduce the impact of unbalanced data to FSVM, improve the classification accuracy.

Abstract: In this paper, we proposed a method for mining mobile users’ Quality of Experience (OoE) model based on weighted LDA. In the recent years, QoE has become an important concept for the quality of networks and services. At present , QoE has attracted the interest of network operators and service providers, because of providing a good QoE service to their customers can satisfy the customers and bring more users. In this paper, we are trying to build up users’ QoE model through topic model, an approach to generate a generative model for data mining. Latent Dirichlet Allocation (LDA) is a feasible and effective algorithm in text modeling. We propose an weighted LDA-based interest model within the modeling framework, and evaluate it on a mobile network users’ behavior extraction system. In this system, we can analyze the users’ behaviors, and build up a vector model for each user through a simple way. Besides, with the help of the topic model, we can get an exact model for users’ QoE, because we can generate the topic model through the vector model. Thus we can get the users’ QoE model, through which we can learn each user’s experience. Then the network operators can provide a better network service for their customers. In the end, we elaborate QoE management requirements for mobile network scenarios, and provide a QoE modeling approach for the mobile network scenarios.

Abstract: Transthoracic impedance (TTI) has been demonstrated to be a potential indicator to monitor the quality of chest compressions (CCs) during cardiopulmonary resuscitation (CPR). However, TTI signals are challenged by noise artifact from multiple sources, such as ventilations and baseline drift. Practically, it is very essential to accurately detect the peak-to-trough of the complicated TTI signals. However, nowadays, there is no method to solve the problem. In this paper, Extrima search with niche technology was used to search the peak-to-trough of TTI signal. We select 2 features to judge the potential peaks and troughs in order to remove the false ones. Besides, we designed a LDA classifier for recognizing the compression and ventilation waves. The experimental results show that this method in this paper can precisely recognize the real peaks and troughs of TTI signals which include some false ones.

Abstract: Twelve vinegars from different production areas in China were evaluated by the Portable Electronic Nose 3 (PEN3), and the data detected by PEN3 were analyzed by principal component analysis (PCA), linear discriminant analysis (LDA) and loadings analysis (LA). The results of PCA and LDA all showed that the electronic nose could clearly discriminate the vinegars of difference production areas, but had very little discrimination on same production area vinegars. The results of LD showed that these sensors, including W1S, W2S, W5S, W2W, W1W, W5C, were appropriate to evaluate and compare the aroma of vinegars, especially W1S, W2S, W5S. Loadings analysis also indicated that these compounds in twelve vinegars had great differences, such as acids, esters, alcohols, alkanes, while aromatics compounds, sulfur-containing compounds and alkenes had some distinctions.

Abstract: KPCA extracting principal component with nonlinear method is an improved PCA. The KPCA can extract the feature set which is more suitable in categorization than the conventional PCA. The method of KFDA is equivalent to KPCA plus LDA. KPCA is first performed and then LDA is used for a second feature extraction in the KPCA-transformed space. The KPCA and KFDA have been got widely used in feature extraction and face recognition. In this paper, the method of KPCA and KFDA is analyzed and their nature is revealed. Finally, the effectiveness of the algorithm is verified using the ORL database.

Abstract: First principle calculations are carried out to investigate the ground state electronic and magnetic properties of G-type antiferromagnetic (AFM) LaCrO₃ compound in orthorhombic phase. The orthorhombic LaCrO₃, a distorted GdFeO₃ – type perovskite exhibits AFM transition at ~290 K. Additionally; this system shows weak ferromagnetism at room temperature, in conjugation with its semiconducting properties. In this paper, we will discuss the onset of weak ferromagnetism and electronic properties of LaCrO₃, implementing generalized gradient approximation for exchange correlation potential. We present results for electronic structure and density of states, including the effect of onsite Coulomb interaction and non-collinear arrangement of Cr³⁺ magnetic ions in G-type AFM geometry.

Abstract: It is difficult to segment instances of object classes accurately unsupervised in images, because of the complexity of structures, inter-class differences, background interference and so on. A multi-scale semantic model method is proposed to overcome the disadvantages existing in most of the relative methods. This method uses generative model to deal with the objects obtained by multi-scale segmentations instead of whole image, and calculates kinds of visual features to mine the topic information of every object. In the segmentation process, a semantic correlative function of every segment block based on KL divergence is built up and minimized to select the object correct regions. Experimental results demonstrate the effectiveness of the proposed method.

Abstract: The electronic band structure, density of state and elastic properties of lead-free perovskite oxide SnTiO3 (ST) were investigated by employing first principles calculation using the Density Functional Theory (DFT) within local density approximation (LDA). The energy band gap was calculated from the separation between the Ti 3d (conduction band) and the maximum of O 2p (valence band). This gives an indirect band gap of 2.36 eV. The elastic constants and their pressure dependence were calculated up to 30 GPa and the independent elastic constants (C11, C12, and C44), bulk modules, B were obtained and analyzed. The results showed that SnTiO3 have a mechanical stability in cubic phase (Pm3m).