THE COMBINATION PREDICTION OF TRANSMEMBRANE REGIONS BASED ON DEMPSTER-SHAFER THEORY OF EVIDENCE

Transmembrane proteins are some special and important proteins in cells. Because of their importance and specificity, the prediction of the transmembrane regions has very important theoretical and practical significance. At present, the prediction methods are mainly based on the physicochemical property and statistic analysis of amino acids. However, these methods are suitable for some environments but inapplicable for other environments. In this paper, the multi-sources information fusion theory has been introduced to predict the transmembrane regions. The proposed method is test on a data set of transmembrane proteins. The results show that the proposed method has the ability of predicting the transmembrane regions as a good performance and powerful tool.

Theory,Method and Application of a State Recognition Based on Theory of Evidence

A new method of state recognition based on the theory of evidence was proposed. By this method, the plausible function which the sample awaiting recognition belongs to each category can be obtained through distance function. For the marginal samples,two or a batch of evidences can be combined and a new plausible function can be obtained by new evidence. Then the categories of samples can be determined according to plausibility function. This method provides a beder reasoning framework. The result proves the rate of recoghition correctness.

Using the Dempster-Shafer Theory of Evidence to Rank Documents

Multi-source information can be utilized collaboratively to improve the performance of information retrieval. To make full use of the document and collection information, this paper introduces a new informa- tion retrieval model that relies on the Dempster-Shafer theory of evidence. Each query-document pair is taken as a piece of evidence for the relevance between a document and a query. The evidence is combined using Dempster＇s rule of combination, and the belief committed to the relevance is obtained. Retrieved documents are then ranked according to the belief committed to the relevance. Several basic probability as- signments are also proposed. Extensive experiments over the Text REtrieval Conference （TREC） test col- lection ClueWeb09 show that the proposed model provides performance similar to that of the Vector Space Model （VSM）. Under certain probability assignments, the proposed model outperforms the VSM by 63% in terms of mean average precision,

AN APPROACH FOR EVIDENCE CLUSTERING USING GENERALIZED DISTANCE

When evidence comes from multiple events they should be handled independently, and it is unknown to which event a piece of evidence is related. In this paper, the problem of clustering all pieces of evidence is analyzed systematically to separate them into subsets for each event on the basis of considering the describing format of evidence and making full use of the distance of evidence. An approach for evidence clustering using distance of evidence is presented based on the criterion for clustering. In the proposed approach, the method which is used to establish the initialization of clus- tering is discussed in detail, called an improved optimal distance. And the centroid vector of evidence and the clustering process are developed respectively to obtain the performance of this novel approach. Finally, an illustrative example shows that this approach is feasible and effective.

DETERMINGING BPA UNDER UNCERTAINTY ENVIRONMENTS AND ITS APPLICATION IN DATA FUSION

Dempster-Shafer (DS) theory of evidence has been widely used in many data fusion ap- plication systems. However, how to determine basic probability assignment, which is the main and the first step in evidence theory, is still an open issue. In this paper, a new method to obtain Basic Probability Assignment (BPA) is proposed based on the similarity measure between generalized fuzzy numbers. In the proposed method, species model can be constructed by determination of the min, average and max value to construct a fuzzy number. Then, a new Radius Of Gravity (ROG) method to determine the similarity measure between generalized fuzzy numbers is used to calculate the BPA functions of each instance. Finally, the efficiency of the proposed method is illustrated by the classi- fication of Iris data.

Land Cover Classification with Multi-source Data Using Evidential Reasoning Approach

Land cover classification is the core of converting satellite imagery to available geographic data.However,spectral signatures do not always provide enough information in classification decisions.Thus,the application of multi-source data becomes necessary.This paper presents an evidential reasoning (ER) approach to incorporate Landsat TM imagery,altitude and slope data.Results show that multi-source data contribute to the classification accuracy achieved by the ER method,whereas play a negative role to that derived by maximum likelihood classifier (MLC).In comparison to the results derived based on TM imagery alone,the overall accuracy rate of the ER method increases by 7.66% and that of the MLC method decreases by 8.35% when all data sources (TM plus altitude and slope) are accessible.The ER method is regarded as a better approach for multi-source image classification.In addition,the method produces not only an accurate classification result,but also the uncertainty which presents the inherent difficulty in classification decisions.The uncertainty associated to the ER classification image is evaluated and proved to be useful for improved classification accuracy.

Remarks on “A new non-specificity measure in evidence theory based on belief intervals”

Two types of uncertainty co-exist in the theory of evidence： discord and non-specificity.From 90s, many mathematical expressions have arisen to quantify these two parts in an evidence.An important aspect of each measure presented is the verification of a coherent set of properties.About non-specificity, so far only one measure verifies an important set of those properties. Very recently, a new measure of non-specificity based on belief intervals has been presented as an alternative measure that quantifies a similar set of properties（Yang et al., 2016）. It is shown that the new measure really does not verify two of those important properties. Some errors have been found in their corresponding proofs in the original publication.

Integrating Machine Learning and Evidential Reasoning for User Profiling and Recommendation

User profiles representing users’preferences and interests play an important role in many applications of personalized recommendation.With the rapid growth of social platforms,there is a critical need for efficient solutions to learn user profiles from the information they shared on social platforms so as to improve the quality of recommendation services.The problem of user profile learning is significantly challenging due to difficulty in handling data from multiple sources,in different formats and often associated with uncertainty.In this paper,we introduce an integrated approach that combines advanced Machine Learning techniques with evidential reasoning based on Dempster-Shafer theory of evidence for user profiling and recommendation.The developed methods for user profile learning and multi-criteria collaborative filtering are demonstrated with experimental results and analysis that show the effectiveness and practicality of the integrated approach.A proposal for extending multi-criteria recommendation systems by incorporating user profiles learned from different sources of data into the recommendation process so as to provide better recommendation capabilities is also highlighted.

Unified Classification and Rejection:A One-versus-all Framework

Classifying patterns of known classes and rejecting ambiguous and novel(also called as out-of-distribution(OOD))inputs are involved in open world pattern recognition.Deep neural network models usually excel in closed-set classification while perform poorly in rejecting OOD inputs.To tackle this problem,numerous methods have been designed to perform open set recognition(OSR)or OOD rejection/detection tasks.Previous methods mostly take post-training score transformation or hybrid models to ensure low scores on OOD inputs while separating known classes.In this paper,we attempt to build a unified framework for building open set classifiers for both classification and OOD rejection.We formulate the open set recognition of K-known-class as a(K+1)-class classification problem with model trained on known-class samples only.By decomposing the K-class problem into K one-versus-all(OVA)binary classification tasks and binding some parameters,we show that combining the scores of OVA classifiers can give(K+1)-class posterior probabilities,which enables classification and OOD rejection in a unified framework.To maintain the closed-set classification accuracy of the OVA trained classifier,we propose a hybrid training strategy combining OVA loss and multi-class cross-entropy loss.We implement the OVA framework and hybrid training strategy on the recently proposed convolutional prototype network and prototype classifier on vision transformer(ViT)backbone.Experiments on popular OSR and OOD detection datasets demonstrate that the proposed framework,using a single multi-class classifier,yields competitive performance in closed-set classification,OOD detection,and misclassification detection.The code is available at https://github.com/zhen-cheng121/CPN_OVA_unified.

The use of evidential belief functions for mineral potential mapping in the Nanling belt, South China

In this study, the evidential belief functions （EBFs） were applied for mapping tungsten polymetallic potential in the Nanling belt, South China. Seven evidential layers （e.g., geological, geochemical, and geophysical） related to tungsten polymetallic deposits were extracted from a multi-source geospatial database. The relationships between evidential layers and the target deposits were quantified using EBFs model. Four EBF maps （belief map, disbelief map, uncertainty map, and plausibility map） are generated by integrating seven evidential layers which provide meaningful interpretations for tungsten polymetallic potential. On the final predictive map, the study area was divided into three target zones of high potential, moderate potential, and low potential areas, among which high potential and moderate potential areas accounted for 17.8% of the total area, containing 81% of the total deposits. To evaluate the success rate accuracy, the receiver operating characteristic （ROC） curves and the area under the curves （AUC） for the belief map were calculated. The area under the curve is 0.81 which indicates that the capability for correctly classifying the areas with existing mineral deposits is satisfactory. The results of this study indicate that the EBFs were effectively used for mapping mineral potential and for managing uncertainties asso- ciated with evidential layers.

A Study on the Model of Fact‑Finding in Criminal Investigation

Fact‑finding,as the foundation of a judicial decision,has been an important consideration in China’s judicial reform.This study introduces the theory of evidence‑based information and falsification methods in the fact‑finding procedure of criminal investigations and proposes a paradigm for fact‑finding using combined pairs of approaches:individual evidence examination and global analysis,the objective basis and subjective perception of fact‑finders,and methods of verification and falsification.The working procedure of the paradigm is illustrated with the objective of making a contribution to the improvement of the existing model of fact‑finding in the criminal justice process.