Peiyu Liu

A method of spam filtering based on weighted support vector machines

The problem of content-based spam filtering on machine learning methods actually is a binary classification. SVMs can separate the data into two categories optimally so SVMs suit to spam filtering. With used into spam filtering, the standard support vector machine involves the minimization of the error function and the accuracy of the SVM is very high, but the degree of misclassification of legitimate emails is high. In order to solve that problem, this paper proposed a method of spam filtering based on weighted support vector machines. Experimental results show that the algorithm can enhance the filtering performance effectively.

An improved Apriori algorithm for association rules of mining

Apriori --the classical association rules mining algorithm is a way to find out certain potential, regular knowledge from the massive ones. But there are two more serious defects in the data mining process. The first needs many times to scan the business database and the second will inevitably produce a large number of irrelevant candidate sets which seriously occupy the system resources. An improved method is introduced on the basic of the defects above. The improved algorithm only scans the database once, at the same time the discrete data and statistics related are completed, and the final one is to prune the candidate item sets according to the minimum supporting degree and the character of the frequent item sets. After analysis, the improved algorithm reduces the system resources occupied and improves the efficiency and quality.

Captopril Pretreatment Produces an Additive Cardioprotection to Isoflurane Preconditioning in Attenuating Myocardial Ischemia Reperfusion Injury in Rabbits and in Humans.

Background. Pretreatment with the angiotensin-converting inhibitor captopril or volatile anesthetic isoflurane has, respectively, been shown to attenuate myocardial ischemia reperfusion (MI/R) injury in rodents and in patients. It is unknown whether or not captopril pretreatment and isoflurane preconditioning (Iso) may additively or synergistically attenuate MI/R injury. Methods and Results. Patients selected for heart valve replacement surgery were randomly assigned to five groups: untreated control (Control), captopril pretreatment for 3 days (Cap3d), or single dose captopril (Cap1hr, 1 hour) before surgery with or without Iso (Cap3d+Iso and Cap1hr+Iso). Rabbit MI/R model was induced by occluding coronary artery for 30 min followed by 2-hour reperfusion. Rabbits were randomized to receive sham operation (Sham), MI/R (I/R), captopril (Cap, 24 hours before MI/R), Iso, or the combination of captopril and Iso (Iso+Cap). In patients, Cap3d+Iso but not Cap1hr+Iso additively reduced postischemic myocardial injury and attenuated postischemic myocardial inflammation. In rabbits, Cap or Iso significantly reduced postischemic myocardial infarction. Iso+Cap additively reduced cellular injury that was associated with improved postischemic myocardial functional recovery and reduced myocardial apoptosis and attenuated oxidative stress. Conclusion. A joint use of 3-day captopril treatment and isoflurane preconditioning additively attenuated MI/R by reducing oxidative stress and inflammation.

The research of an improved information gain method using distribution information of terms

The inadequacy of the information gain is taken into account the situation that the term does not appear. But, in this paper, by analyzing the distribution information of terms, we find if the value of Distribution Information inside a Class of the term becomes large, the distribution of the term inclines to imbalance, and if there is high imbalance of the term, the Distribution Information among Classes will tend to a smaller value. Therefore, the Distribution Information inside a Class and Distribution Information among Classes are introduced to this paper to reduce the effect of the term does not appear, and improve the traditional information gain. After experimental verification, the improved algorithm (GDI) has a better performance than traditional feature selection algorithm in some fields, such as the Information Gain.

A Self-Adaptive Fuzzy c-Means Algorithm for Determining the Optimal Number of Clusters

For the shortcoming of fuzzy c-means algorithm (FCM) needing to know the number of clusters in advance, this paper proposed a new self-adaptive method to determine the optimal number of clusters. Firstly, a density-based algorithm was put forward. The algorithm, according to the characteristics of the dataset, automatically determined the possible maximum number of clusters instead of using the empirical rule n and obtained the optimal initial cluster centroids, improving the limitation of FCM that randomly selected cluster centroids lead the convergence result to the local minimum. Secondly, this paper, by introducing a penalty function, proposed a new fuzzy clustering validity index based on fuzzy compactness and separation, which ensured that when the number of clusters verged on that of objects in the dataset, the value of clustering validity index did not monotonically decrease and was close to zero, so that the optimal number of clusters lost robustness and decision function. Then, based on these studies, a self-adaptive FCM algorithm was put forward to estimate the optimal number of clusters by the iterative trial-and-error process. At last, experiments were done on the UCI, KDD Cup 1999, and synthetic datasets, which showed that the method not only effectively determined the optimal number of clusters, but also reduced the iteration of FCM with the stable clustering result.

Targeting lncRNAs for cardiovascular therapeutics in coronary artery disease

Long non-coding RNAs (lncRNAs) could regulate gene expression at posttranscriptional state, and might be a new therapeutic option in human disease. Mounting studies in the past few years have documented important functions for lncRNAs in coronary artery diseases (CAD). Among these lncRNAs, H19, MIAT, ANRIL, lincRNA-p21, AC100865.1, OTTHUMT00000387022, NONHSAT112178, Novlnc6, MALAT1, LIPCAR and SRA have been proposed to be novel regulators and/or biomarkers of CAD, and involved in the initiation and progression of the pathology. So, in this review, we summarize the current understandings of these lncRNAs, mainly focusing on the known biological functions and its underlying molecular mechanisms. Finally, we highlight the increasing evidences for the importance of lncRNA in CAD, which indicate lncRNAs as potential diagnostic markers and/or valuable therapeutic targets for the diseases.

An improved fuzzy C-means clustering algorithm based on simulated annealing

Fuzzy C-means clustering algorithm (FCM) is a widely used clustering algorithm, however it has its drawbacks: the initial number of clusters needs to be determined by the manual control according to the prior knowledge; the objective function ignores the disequilibrium problems among the sample attribute data. In view of these problems, this paper proposes a sample weighted FCM algorithm based on simulated annealing algorithm. It uses the simulated annealing algorithm which has an excellent ability of seeking global optimal solution to calculate the initial value of the number of clusters and makes certain weighting process on the clustering center function and the objective function. The experiment results show that this proposed algorithm has better classification accuracy and classification accuracy rate compared with FCM algorithm and the common sample weighted FCM clustering algorithms. Meanwhile, this algorithm needs not to be determined the initial value of clusters manually. The improved algorithm possesses the superiority and the actual application value.

Magnetotactic Coccus Strain SHHC-1 Affiliated to Alphaproteobacteria Forms Octahedral Magnetite Magnetosomes

Magnetotactic bacteria (MTB) are morphologically and phylogenetically diverse prokaryotes. They can form intracellular chain-assembled magnetite (Fe3O4) or greigite (Fe3S4) nanocrystals each enveloped by a lipid bilayer membrane called a magnetosome. Magnetotactic cocci have been found to be the most abundant morphotypes of MTB in various aquatic environments. However, knowledge on magnetosome biomineralization within magnetotactic cocci remains elusive due to small number of strains that have been cultured. By using a coordinated fluorescence and scanning electron microscopy method, we discovered a unique magnetotactic coccus strain (tentatively named SHHC-1) in brackish sediments collected from the estuary of Shihe River in Qinhuangdao city, eastern China. It phylogenetically belongs to the Alphaproteobacteria class. Transmission electron microscopy analyses reveal that SHHC-1 cells formed many magnetite-type magnetosomes organized as two bundles in each cell. Each bundle contains two parallel chains with smaller magnetosomes generally located at the ends of each chain. Unlike most magnetotactic alphaproteobacteria that generally form magnetosomes with uniform crystal morphologies, SHHC-1 magnetosomes display a more diverse variety of crystal morphology even within a single cell. Most particles have rectangular and rhomboidal projections, whilst others are triangular, or irregular. High resolution transmission electron microscopy observations coupled with morphological modeling indicate an idealized model—elongated octahedral crystals, a form composed of eight {111} faces. Furthermore, twins, multiple twins and stack dislocations are frequently observed in the SHHC-1 magnetosomes. This suggests that biomineralization of strain SHHC-1 magnetosome might be less biologically controlled than other magnetotactic alphaproteobacteria. Alternatively, SHHC-1 is more sensitive to the unfavorable environments under which it lives, or a combination of both factors may have controlled the magnetosome biomineralization process within this unique MTB.

Improved Biogeography-Based Optimization Based on Affinity Propagation

To improve the search ability of biogeography-based optimization (BBO), this work proposed an improved biogeography-based optimization based on Affinity Propagation. We introduced the Memetic framework to the BBO algorithm, and used the simulated annealing algorithm as the local search strategy. MBBO enhanced the exploration with the Affinity Propagation strategy to improve the transfer operation of the BBO algorithm. In this work, the MBBO algorithm was applied to IEEE Congress on Evolutionary Computation (CEC) 2015 benchmarks optimization problems to conduct analytic comparison with the first three winners of the CEC 2015 competition. The results show that the MBBO algorithm enhances the exploration, exploitation, convergence speed and solution accuracy and can emerge as the best solution-providing algorithm among the competing algorithms.

A logical paragraph division based on semantic characteristics and its application

In this paper, we introduced a new matching method based on logic-centered paragraphs. This method is built on the concept dictionary, in this method, the paragraphs which have the same meaning will be clustered by analyzing the logical concept of the text to be classified, and establish the logical paragraphs on the basis of the division method of logical levels. Then put the text to be classified in the right classification, which considered the contribution to the theme of different paragraphs in the text. At the same time, in order to solve problem of synonyms and polysemy in the texts to be classified, we introduced the expansion of the synonyms concept and related words. Experimental results show that this method can improve the effectiveness of classification, and a higher accuracy rate can be obtained in content flitting.