Support vector machines for credit scoring and discovery of significant features

The assessment of risk of default on credit is important for financial institutions. Logistic regression and discriminant analysis are techniques traditionally used in credit scoring for determining likelihood to default based on consumer application and credit reference agency data. We test support vector machines against these traditional methods on a large credit card database. We find that they are competitive and can be used as the basis of a feature selection method to discover those features that are most significant in determining risk of default.     

Pap smear classification using combination of global significant value,texture statistical features and time series features

Multimedia Tools and Applications - Pap smear test is routinely used today to diagnose cervical cancer. In the last 50 years, this simple test has saved millions of women’s lives....     

A novel neural-genetic algorithm to find the most significant combination of features in digital mammograms

Digital mammography is one of the most suitable methods for early detection of breast cancer. It uses digital mammograms to find suspicious areas containing benign and malignant microcalcifications. However, it is very difficult to distinguish benign and malignant microcalcifications. This is reflected in the high percentage of unnecessary biopsies that are performed and many deaths caused by late detection or misdiagnosis. A computer based feature selection and classification system can provide a second opinion to the radiologists in assessment of microcalcifications. The research in this paper proposes a neural-genetic algorithm for feature selection to classify microcalcification patterns in digital mammograms. It aims to develop a step-wise algorithm to find the best feature set and a suitable neural architecture for microcalcification classification. The obtained results show that the proposed algorithm is able to find an appropriate feature subset, which also produces a high classification rate.     

Formal validation of domain-specific languages with derived features and well-formedness constraints

Despite the wide range of existing tool support, constructing a design environment for a complex domain-specific language (DSL) is still a tedious task as the large number of derived features and well-formedness constraints complementing the domain metamodel necessitate special handling. Such derived features and constraints are frequently defined by declarative techniques (such graph patterns or OCL invariants). However, for complex domains, derived features and constraints can easily be formalized incorrectly resulting in inconsistent, incomplete or ambiguous DSL specifications. To detect such issues, we propose an automated mapping of EMF metamodels enriched with derived features and well-formedness constraints captured as graph queries in EMF-IncQuery or (a subset of) OCL invariants into an effectively propositional fragment of first-order logic which can be efficiently analyzed by back-end reasoners. On the conceptual level, the main added value of our encoding is (1) to transform graph patterns of the EMF-IncQuery framework into FOL and (2) to introduce approximations for complex language features (e.g., transitive closure or multiplicities) which are not expressible in FOL. On the practical level, we identify and address relevant challenges and scenarios for systematically validating DSL specifications. Our approach is supported by a tool, and it will be illustrated on analyzing a DSL in the avionics domain. We also present initial performance experiments for the validation using Z3 and Alloy as back-end reasoners.     

Appropriate inferences of data dependencies in relational databases

We study inference systems for the combined class of functional and full hierarchical dependencies in relational databases. Two notions of implication are considered: the original notion in which a dependency is implied by a given set of dependencies and the underlying set of attributes, and the alternative notion in which a dependency is implied by a given set of dependencies alone. The first main result establishes a finite axiomatisation for the original notion of implication which clarifies the role of the complementation rule in the combined setting. In fact, we identify inference systems that are appropriate in the following sense: full hierarchical dependencies can be inferred without use of the complementation rule at all or with a single application of the complementation rule at the final step of the inference only; and functional dependencies can be inferred without any application of the complementation rule. The second main result establishes a finite axiomatisation for the alternative notion of implication. We further show how inferences of full hierarchical dependencies can be simulated by inferences of multivalued dependencies, and vice versa. This enables us to apply both of our main results to the combined class of functional and multivalued dependencies. Furthermore, we establish a novel axiomatisation for the class of non-trivial functional dependencies.     

Recognition and analysis of cell nuclear phases for high-content screening based on morphological features

Automated analysis of molecular images has increasingly become an important research in computational life science. In this paper some new and efficient algorithms for recognizing and analyzing cell phases of high-content screening are presented. The conceptual frameworks are based on the morphological features of cell nuclei. The useful preprocessing includes: smooth following and linearization; extraction of morphological structural points; shape recognition based morphological structure; issue of touching cells for cell separation and reconstruction. Furthermore, the novel detecting and analyzing strategies of feed-forward and feed-back of cell phases are proposed based on gray feature, cell shape, geometrical features and difference information of corresponding neighbor frames. Experiment results tested the efficiency of the new method.     

Incorporating multiple genomic features with the utilization of interacting domain patterns to improve the prediction of protein-protein interactions

Protein-protein interaction (PPI) networks play an outstanding role in the organization of life. Parallel to the growth of experimental techniques on determining PPIs, the emergence of computational methods has greatly accelerated the time needed for the identification of PPIs on a wide genomic scale. Although experimental approaches have limitations that can be complemented by the computational methods, the results from computational methods still suffer from high false positive rates which contribute to the lack of solid PPI information. Our study introduces the PPI-Filter; a computational framework aimed at improving PPI prediction results. It is a post-prediction process which involves filtration, using information based on three different genomic features; (i) gene ontology annotation (GOA), (ii) homologous interactions and (iii) protein families (PFAM) domain interactions. In the study, we incorporated a protein function prediction method, based on interacting domain patterns, the protein function predictor or PFP (), for the purpose of aiding the GOA. The goal is to improve the robustness of predicted PPI pairs by removing the false positive pairs and sustaining as much true positive pairs as possible, thus achieving a high confidence level of PPI datasets. The PPI-Filter has been proven to be applicable based on the satisfactory results obtained using signal-to-noise ratio (SNR) and strength measurements that were applied on different computational PPI prediction methods.     

Exact inferences in a neural implementation of a hidden Markov model

From first principles, we derive a quadratic nonlinear, first-order dynamical system capable of performing exact Bayes-Markov inferences for a wide class of biologically plausible stimulus-dependent patterns of activity while simultaneously providing an online estimate of model performance. This is accomplished by constructing a dynamical system that has solutions proportional to the probability distribution over the stimulus space, but with a constant of proportionality adjusted to provide a local estimate of the probability of the recent observations of stimulus-dependent activity-given model parameters. Next, we transform this exact equation to generate nonlinear equations for the exact evolution of log likelihood and log-likelihood ratios and show that when the input has low amplitude, linear rate models for both the likelihood and the log-likelihood functions follow naturally from these equations. We use these four explicit representations of the probability distribution to argue that, in contrast to the arguments of previous work, the dynamical system for the exact evolution of the likelihood (as opposed to the log likelihood or log-likelihood ratios) not only can be mapped onto a biologically plausible network but is also more consistent with physiological observations.     

A unified sample preparation protocol for proteomic and genomic profiling of cervical swabs to identify biomarkers for cervical cancer screening

Cervical cancer screening is ideally suited for the development of biomarkers due to the ease of tissue acquisition and the well-established histological transitions. Furthermore, cell and biologic fluid obtained from cervix samples undergo specific molecular changes that can be profiled. However, the ideal manner and techniques for preparing cervical samples remains to be determined. To address this critical issue a patient screening protein and nucleic acid collection protocol was established. RNAlater was used to collect the samples followed by proteomic methods to identify proteins that were differentially expressed in normal cervical epithelial versus cervical cancer cells. Three hundred ninety spots were identified via 2-D DIGE that were expressed at either higher or lower levels (>three-fold) in cervical cancer samples. These proteomic results were compared to genes in a cDNA microarray analysis of microdissected neoplastic cervical specimens to identify overlapping patterns of expression. The most frequent pathways represented by the combined dataset were: cell cycle: G2/M DNA damage checkpoint regulation; aryl hydrocarbon receptor signaling; p53 signaling; cell cycle: G1/S checkpoint regulation; and the ER stress pathway. HNRPA2B1 was identified as a biomarker candidate with increased expression in cancer compared to normal cervix and validated by Western blot.     

An effective HIV-1 integrase inhibitor screening platform: Rationality validation of drug screening,conformational mobility and molecular recognition analysis for PFV integrase complex with viral DNA

As an important target for the development of novel anti-AIDS drugs, HIV-1 integrase (IN) has been widely concerned. However, the lack of a complete accurate crystal structure of HIV-1 IN greatly blocks the discovery of novel inhibitors. In this work, an effective HIV-1 IN inhibitor screening platform, namely PFV IN, was filtered from all species of INs. Next, the 40.8% similarity with HIV-1 IN, as well as the high efficiency of virtual screening and the good agreement between calculated binding free energies and experimental ones all proved PFV IN is a promising screening platform for HIV-1 IN inhibitors. Then, the molecular recognition mechanism of PFV IN by its substrate viral DNA and six naphthyridine derivatives (NRDs) inhibitors was investigated through molecular docking, molecular dynamics simulations and water-mediated interactions analyses. The functional partition of NRDs IN inhibitors could be divided into hydrophobic and hydrophilic ones, and the Mg²⁺ ions, water molecules and conserved DDE motif residues all interacted with the hydrophilic partition, while the bases in viral DNA and residues like Tyr212, Pro214 interacted with the hydrophobic one. Finally, the free energy landscape (FEL) and cluster analyses were performed to explore the molecular motion of PFV IN-DNA system. It is found that the association with NRDs inhibitors would obviously decrease the motion amplitude of PFV IN-DNA, which may be one of the most potential mechanisms of IN inhibitors. This work will provide a theoretical basis for the inhibitor design based on the structure of HIV-1 IN.     

Automated screening of congestive heart failure using variational mode decomposition and texture features extracted from ultrasound images

Heart is an important and hardest working muscular organ of the human body. Inability of the heart to restore normal perfusion to the entire body refers to cardiac failure, which then with symptoms results in manifestation of congestive heart failure (CHF). Impairment in systolic function associated with chronic dilation of left ventricle is referred as dilated cardiomyopathy (DCM). The clinical examination, surface electrocardiogram (ECG), chest X-ray, blood markers and echocardiography play major role in the diagnosis of CHF. Though the ECG manifests chamber enlargement changes, it does not possess sensitive marker for the diagnosis of DCM, whereas echocardiographic assessment can effectively reveal the presence of asymptomatic DCM. This work proposes an automated screening method for classifying normal and CHF echocardiographic images affected due to DCM using variational mode decomposition technique. The texture features are extracted from variational mode decomposed image. These features are selected using particle swarm optimization and classified using support vector machine classifier with different kernel functions. We have validated our experiment using 300 four-chamber echocardiography images (150: normal, 150: CHF) obtained from 50 normal and 50 CHF patients. Our proposed approach yielded maximum average accuracy, sensitivity and specificity of 99.33%, 98.66% and 100%, respectively, using ten features. Thus, the developed diagnosis system can effectively detect CHF in its early stage using ultrasound images and aid the clinicians in their diagnosis.

     

Least significant bit evaluation of arithmetic expressions in single-precision

Single-precision floatingpoint computations may yield an arbitrary false result due to cancellation and rounding errors. This is true even for very simple, structured arithmetic expressions such as Horner's scheme for polynomial evaluation. A simple procedure will be presented for fast calculation of the value of an arithmetic expression to least significant bit accuracy in single precision computation. For this purpose in addition to the floating-point arithmetic only a precise scalar product (cf. [2]) is required. If the initial floatingpoint approximation is not too bad, the computing time of the new algorithm is approximately the same as for usual floating-point computation. If not, the essential progress of the presented algorithm is that the inaccurate approximation is recognized and corrected. The algorithm achieves high accuracy, i.e. between the left and the right bound of the result there is at most one more floating-point number. A rigorous estimation of all rounding errors introduced by floating-point arithmetic is given for general triangular linear systems. The theorem is applied to the evaluation of arithmetic expressions.     

Zones of low entropy in genomic sequences.

We consider the problem of detecting regions having low entropy in DNA sequences, which is a particular case of searching for dos-DNA zones. The entropy is measured in linear time as the number of distinct segments occurring in the regions. As a consequence, we are able to determine regions containing a small number of repetitions of long segments or a large number of repetitions of short segments. The method provides an index on sequences that is applied to compare them without any alignment. Comparisons extract regions having similar combinatorial features that would not have been found by standard alignment methods. The present methodology is applied to Saccharomyces cerevisiae yeast chromosomes to show what the approach is able to produce.     

Emergence of genomic self-similarity in location independent representations

A key property for predicting the effectiveness of stochastic search techniques, including evolutionary algorithms, is the existence of a positive correlation between the form and the quality of candidate solutions. In this paper we show that when the ordering of genomic symbols in a genetic algorithm is completely independent of the fitness function and therefore free to evolve along with the candidate solutions it encodes, the resulting genomes self-organize into self-similar structures that favor this key stochastic search property.     

Microoptical characterization of piezoelectric vibratory microinjections in drosophila embryos for genome-wide RNAi screen

In this paper, we study the effect of acoustic agitation on the penetration force for microinjections in Drosophila embryos for genome-wide RNA interference (RNAi) screens, using an integrated optical MEMS force encoder for in vivo characterization of the dynamic penetration forces. Two modes of operation are investigated. In the first mode of operation, the injector is brought into contact and acts on the embryo with a fixed force, and the vibration amplitude of the microinjector is increased till penetration occur. We observed a linear decrease in the penetration force of 1.6 /spl mu/N with every 0.1 m/s tip velocity increase. In the second mode of operation, the vibration amplitude is kept constant and the injector is pushed into the embryo until penetration. We simulate the optical force encoder eigenmodes and measure the injection force over the frequency range from 0 to 16 kHz with actuation voltages up to 150V. Among the eight encoder eigenmodes with resonant frequency up to 16 kHz, the longitudinal vibration along the injector is shown to dominate the force reduction at 14 kHz. Two other modes, both involving significant out-of-plane injector motion, reduce the penetration force by 52% around 3.1 kHz. The average penetration force is calculated based on injections into multiple embryos for each experimental condition. For each microinjection, the peak (or average) penetration force can be derived from the peak (or average) relative displacement of the two gratings upon penetration. The achieved minimum peak penetration force was 15.6 /spl mu/N (/spl sim/29.7% of the static penetration force), while the minimum average penetration force was 2.7 /spl mu/N (5.1% of the static penetration force).     

图像最低有效位信息隐藏分析

信息隐藏检测技术是近年来一个新兴的研究领域,用统计实证方法对最低有效位替换数字图像的数据特性进行研究分析,利用S-Tools对相同载体嵌入不同格式图像、不同载体嵌入相同格式图像的实例的分析,说明利用差分直方图可以检测图像最低有效位的信息隐藏,并通过仿真说明方法的有效性.     

Role of identifiability in validation of compartmental models

State and parameter estimation algorithms can $\text{validity}$ be applied only to a system which is observable, controllable and identifiable. The roles of observability, controllability and especially $\text{identifiability}$ are considered in achieving a system having unique structure and unique set of parameters. Analysis is simplified considerably by restricting the class of system to the classical linear $\text{compartmental}$ system.     

Efficient validation of mobile transactions in wireless environments

In broadcast environments, the limited bandwidth of the upstream communication channel from the mobile clients to the server bars the application of conventional concurrency control protocols. In this paper, we propose a new variant of the optimistic concurrency control (OCC) protocol that is suitable for broadcast environments. At the server, forward validation of a transaction is done against currently running transactions, including mobile transactions and server transactions. At the mobile clients, partial backward validation of a transaction is done against committed transactions at the beginning of every broadcast cycle. Upon completion of execution, read-only mobile transactions can be validated and committed locally and update mobile transactions are sent to the server for final validation. These update transactions have a better chance of commitment because they have gone through the partial backward validation. In addition to the nice properties of conventional OCC protocols, this protocol provides autonomy between the mobile clients and the server with minimum upstream communication, which is a desirable feature to the scalability of applications running in broadcast environments. This protocol is able to process both update transactions and read-only transactions at the mobile clients at low space and processing overheads.     

JAVA和XML Schema在数据有效性验证中的应用

数据的有效性验证是对输入数据正误的判断，分析了几种不同的数据验证方法，提出了新的验证方法即将XML Schema用于JAVA程序的数据有效性验证的可行性，并部分实现这种方案。