Similarity Calculation Method Research Articles

Social web video clustering based on multi-modal and clustering ensemble

Web videos are rich resources for people to satisfy their information and entertainment needs. Previous studies have applied clustering methods by using textual information of videos tagged by the up-loaders to perform web video categorization, which helps users easily find videos that they really want and then increase the user’s satisfaction. However, web video categorization remains a challenging task due to the difficulties in accuracy measuring the semantic relation between terms in videos. In this paper, a novel framework for social web video clustering is proposed by improving the similarity calculation method of web videos and using the clustering ensemble. It consists of the following steps: 1) A new semantic based on Vector Space Model (VSM) is defined by considering the semantic relation of terms obtained from the lexical reference system (WordNet). 2) Word2vec is used to capture the continuous vectors as semantic information in the form of vector set of terms in a document. 3) The comprehensive extension of Semantic VSM by utilizing the Normalized Google Distance is presented. 4) The linear combining function is embodied to combine the similarity based on the optimal values of the parameter to control the weights of models before applying them to clustering paradigms and the Clustering Ensemble is employed to integrate the results of each clustering with Must-Link constraint. Experimental evaluations on real-world social web video datasets demonstrate that the proposed method effectively facilitates the clustering and achieves promising performance.

Promotive Effect of Psychological Intervention on English Vocabulary Teaching Based on Hybrid Collaborative Recommender Technology

Psychological changes will often play a huge influence on the teaching effects in the English vocabulary learning and teaching process. To further explore the relationship of psychological regulation with English vocabulary teaching, the hybrid collaborative recommender technology may be underlying to propose a hybrid algorithm for studying how the English vocabulary teaching effect is subjected to the psychological emotion factors as parameters added in the original similarity calculation method in the psychological regulation environment where the emotion is the core of study. In the end, the paper conducts the questionnaire survey among 300 students. Test data show that positive emotion acts as the catalyst of the English vocabulary teaching effect, having basically reached the expected goal of study.

Research on Hot Topics Discovery Based on Short Texts of Online Reviews

The two crucial issues for hot topics discovery based on online reviews are that the sparsity of short text features and the “long tail” phenomenon of hot topics. This paper focuses on these two key issues, and proposes an improved similarity calculation method to calculate the similarity of short texts, and a novel clustering algorithm based on the time factor and dynamic adjustment of comparison times to automatically discard a large number of outliers. Moreover, the validity and advancement of the new method are presented by comparative experiments using real data sets.

Improvement of the calculation method of cyclone dust collectors

The object of research is a method for calculating the efficiency of cyclones using the magnitude of the velocity of transverse pulsations of a turbulent gas flow in the inlet of the cyclone. One of the most problematic issues when creating a new cyclonic technique for gas dust cleaning is the need to carry out a fairly large amount of work – modeling cyclones in laboratory conditions on model cyclones, taking into account the large-scale transition, etc. Therefore, the development of the most accurate analytical method for calculating the efficiency of cyclones is an important problem since it makes it possible to quickly and accurately determine all the necessary cyclone parameters for specific production conditions.During the study, patterns of dust particle deposition in the approximation of the «diffusion boundary layer» were used, which made it possible to calculate with high accuracy the efficiency of cleaning industrial dust in cyclones of various designs. At the same time, the time and the amount of experimental work are also significantly reduced in the development of new types of cyclones or in their modernization or selection.The most accurate dependencies are obtained for evaluating the most important parameters of dust cleaning efficiency in various cyclones. This is due to the fact that the method for calculating cyclones proposed in the work has a number of features, in particular, the development of new dependences for calculating the cut-off diameter «dh=50», the inertial capture coefficient of dust particles «E», and others. It is also proposed to use the average cross section the inlet pipe of the cyclone of the velocity of transverse pulsations of the turbulent gas flow and the new interpolation dependence of the coefficient of inertial capture of dust particles «E» at the potential flow regime in the inlet pipe of the cyclone x machines. Thanks to this, it is possible to obtain accurate values of the indicators of cleaning efficiency in cyclones of various design parameters. Compared with similar well-known calculation methods, this provides such advantages as accuracy of calculations, speed of execution and reliability of the results.

Determination of the technical state of buildings and constructions after force and temperature influences

The object of research is the technical condition and bearing capacity of reinforced concrete structures of buildings and structures after power and high-temperature influences. Calculation methods, which are recommended by the current regulatory documents of Ukraine, do not always allow to correctly predict the growth of structural deformations and assess the real stock of bearing capacity.One of the most problematic places is the calculation of structures working under forced displacements of supports and/or possible high-temperature influences. The situation is further aggravated by the fact that calculations are carried out, as a rule, using non-deformable schemes.Strengthening of structures of buildings that have suffered damage after various impacts is usually carried out using metal elements. At the same time, the main thing remains to perform structural analysis for the justified purpose of the sections of reinforcing elements.In the course of the study, various methods were used, primarily modeling the operation of structures using the finite element method and modern computer systems. This is due to the fact that the proposed method for solving the problem has a number of features, in particular, it allows determining the distribution of forces in the building elements after changing the stiffness characteristics or introducing additional core elements into the design scheme. In the course of solving the problem, the appearance and development of cracks is modeled by changing the stiffness characteristics of the elements.The efforts that could have arisen in the elements of the building and reinforcement are obtained. Thanks to this, it is possible to make decisions about the possibility of further operation, reinforcement or replacement of structures. Changing the conditions of consolidation is considered as action on the part of the foundation. Compared with similar well-known calculation methods, this approach makes it possible to predict changes in the technical condition over time, that is, taking into account changes in fixing conditions and characteristics of building materials will allow a more reasonable approach to assessing the stress-strain state and the residual life of the structure or structure as a whole.

Method of Process Similarity Analysis Based on Longest Sub-similar Sequence Set

The traditional process similarity calculation method has a large amount of calculation and the calculation process is cumbersome. For the case of long process routes and long codes, the calculation difficulty and running time increase rapidly, and the operation effect is not good. In this paper, a process similarity analysis algorithm based on the longest sub-similar sequence set is proposed. Compared with the traditional calculation method, this algorithm has the advantages of high operating efficiency and strong adaptability. Further considering the influence of continuous similar process sequences, the penalty factor is introduced to improve the algorithm, so that the algorithm is more perfect and reasonable. This algorithm has significant application value in modern manufacturing systems.

Research on Text Mining of Railway Safety Supervisors Performance Based on BiLSTM and CRF

In the era of big data with unstructured data, the potential value of unstructured data mining is an effective way to solve the industry problem. Based on a large number of data recorded from the actual text of the performance plan of railway security supervisors, this paper proposes a text similarity calculation method based on text mining technology to calculate whether the personnel performance plan matches the reality. First, the text Named Entity Recognition(NER) is realized by using the combination of Bi-directional Long Short Term Memory (BiLSTM) and the Conditional Random Fields (CRF). Then based on the concept similarity calculation of HowNet, we designed the method of calculating the similarity between the same named entity. After that, different weights are given according to the importance of named entity, and a method of computing similarity between named entities is designed.. Finally, according to the given threshold, to obtain whether planning and realistic matching, and the actual personnel working situation, provide the basis for personnel assessment for the management personnel.

A novel next new point-of-interest recommendation system based on simulated user travel decision-making process

Abstract POI(point of interest) recommendation systems have been widely investigated in recent years. Currently, most POI recommendation systems only recommend POIs that may be visited by users in the future, and rarely consider next new POI recommendation based on the current time and the current location of a particular user. In fact, next new POI recommendation problem is more challenging for the reason that multiple factors associated with both POIs and users need to be comprehensively incorporated in a unified recommendation system. In this paper, we design a novel and effective next new POI recommendation system. Our system simulates a user’s travel decision-making process by weighing two important factors that affect a user’s travel decision: preference factors and geographic factors. First, we use tensor to model user’s check-in history and dynamically predict user preferences. Then, in order to characterize the influence of geographic factor on individual users, we designed a personalized user similarity calculation method and fitted curves for the target user to reflect the relationship between travel distance and travel probability. Finally, a recommendation list is generated by combining the effects of these two factors on a particular user. Compared with the state-of-the-art POI recommendation approach, the experimental results demonstrate that our system achieves much better performance.

A security risk plan search assistant decision algorithm using deep neural network combined with two-stage similarity calculation

In view of the nonlinearity and uncertainty of safety accident risk assessment, firstly, based on the deep neural network, the training criterion of the network is changed, and the triplet convolutional neural network with the similarity measure as the cost function is proposed. The inactive multi-scale set features are extracted from them, so that the semantic features obtained by learning are suitable for security risk image retrieval. In the image retrieval application, the training samples of the retrieved data set are not enough to train a large network, and the innovative application of migration learning to security risk image retrieval proposes to train the network with data sets similar to the retrieved data sets. Then based on the traditional nearest neighbor algorithm, this paper proposes a case similarity calculation method based on two-dimensional structure of structural similarity and attribute similarity, input the characteristic attribute value of the current emergency event, and conduct similar case retrieval. The final calculation returns the historical case and its solution that the user is most similar to the currently entered incident feature. The experiment proves that the maximum relative error between the output of the network and the expected output value is 5.17%, and the minimum relative error is 1.38%, which has high accuracy.

Improving Semantic Similarity with Cross-Lingual Resources: A Study in Bangla—A Low Resourced Language

Semantic similarity is a long-standing problem in natural language processing (NLP). It is a topic of great interest as its understanding can provide a look into how human beings comprehend meaning and make associations between words. However, when this problem is looked at from the viewpoint of machine understanding, particularly for under resourced languages, it poses a different problem altogether. In this paper, semantic similarity is explored in Bangla, a less resourced language. For ameliorating the situation in such languages, the most rudimentary method (path-based) and the latest state-of-the-art method (Word2Vec) for semantic similarity calculation were augmented using cross-lingual resources in English and the results obtained are truly astonishing. In the presented paper, two semantic similarity approaches have been explored in Bangla, namely the path-based and distributional model and their cross-lingual counterparts were synthesized in light of the English WordNet and Corpora. The proposed methods were evaluated on a dataset comprising of 162 Bangla word pairs, which were annotated by five expert raters. The correlation scores obtained between the four metrics and human evaluation scores demonstrate a marked enhancement that the cross-lingual approach brings into the process of semantic similarity calculation for Bangla.

TSS: Temporal similarity search measure for heterogeneous information networks

Many real-world phenomena can be modeled as networked systems. Some of these systems consist of heterogeneous nodes and edges. Similarity search is a fundamental operation in network systems, which is a basis for various applications such as link prediction and recommendation. This manuscript introduces a temporal similarity measure for heterogeneous networks. The proposed metric is based on metapath strategy and considers time of the interaction. We provide detailed properties of the proposed temporal similarity measures. Our experimental results on a number of real heterogeneous social networks show that incorporating time in the computation of the similarity significantly improves the performance as compared to the state-of-the-art similarity computation methods.

QoS Prediction for Service Recommendation with Deep Feature Learning in Edge Computing Environment

Along with the popularity of intelligent services and mobile services, service recommendation has become a key task, especially the task based on quality-of-service (QoS) in edge computing environment. Most existing service recommendation methods have some serious defects, and cannot be directly adopted in edge computing environment. For example, most of existing methods cannot learn deep features of users or services, but in edge computing environment, there are a variety of devices with different configurations and different functions, and it is necessary to learn deep features behind those complex devices. In order to fully utilize hidden features, this paper proposes a new matrix factorization (MF) model with deep features learning, which integrates a convolutional neural network (CNN). The proposed mode is named Joint CNN-MF (JCM). JCM is capable of using the learned deep latent features of neighbors to infer the features of a user or a service. Meanwhile, to improve the accuracy of neighbors selection, the proposed model contains a novel similarity computation method. CNN learns the neighbors features, forms a feature matrix and infers the features of the target user or target service. We conducted experiments on a real-world service dataset under a batch of cases of data densities, to reflect the complex invocation cases in edge computing environment. The experimental results verify that compared to counterpart methods, our method can consistently achieve higher QoS prediction results.

Group Formation Basd on SATC-ALO and SOM Neural Network

Firstly, the problem of group-air grouping is analyzed to introduce the aircraft attribute grouping model and aircraft fuel consumption grouping model. Then, SATC-ALO optimized by Chaos optimization algorithm and Tournament Selection strategy and SOM neural network are used to solve the formation grouping model. Finally, comparative experiments of similarity calculation method and formation grouping method were performed with 50 groups of data. The experimental results show that hybrid method is superior to Euclidean distance method. SATC-ALO algorithm has the highest grouping accuracyand meets the real-time requirements. However, the number of groups needs to be specified in advance. The accuracy of SOM neural network grouping is slightly lower than SATC-ALO algorithm, but the grouping time is lower than SATC-ALO algorithm, and there is no need to specify the number of groups. Both SOM neural network and SATC-ALO algorithm can perfectly solve the problem of group-air grouping and have practical application value.

Ligand-Based Virtual Screening Using Graph Edit Distance as Molecular Similarity Measure.

Extended reduced graphs provide summary representations of chemical structures using pharmacophore-type node descriptions to encode the relevant molecular properties. Commonly used similarity measures using reduced graphs convert these graphs into 2D vectors like fingerprints, before chemical comparisons are made. This study investigates the effectiveness of a graph-only driven molecular comparison by using extended reduced graphs along with graph edit distance methods for molecular similarity calculation as a tool for ligand-based virtual screening applications, which estimate the bioactivity of a chemical on the basis of the bioactivity of similar compounds. The results proved to be very stable and the graph editing distance method performed better than other methods previously used on reduced graphs. This is exemplified with six publicly available data sets: DUD-E, MUV, GLL&GDD, CAPST, NRLiSt BDB, and ULS-UDS. The screening and statistical tools available on the ligand-based virtual screening benchmarking platform and the RDKit were also used. In the experiments, our method performed better than other molecular similarity methods which use array representations in most cases. Overall, it is shown that extended reduced graphs along with graph edit distance is a combination of methods that has numerous applications and can identify bioactivity similarities in a structurally diverse group of molecules.

Towards Semantic Sensor Data: An Ontology Approach.

In order to optimize intelligent applications driven by various sensors, it is vital to properly interpret and reuse sensor data from different domains. The construction of semantic maps which illustrate the relationship between heterogeneous domain ontologies plays an important role in knowledge reuse. However, most mapping methods in the literature use the literal meaning of each concept and instance in the ontology to obtain semantic similarity. This is especially the case for domain ontologies which are built for applications with sensor data. At the instance level, there is seldom work to utilize data of the sensor instances when constructing the ontologies’ mapping relationship. To alleviate this problem, in this paper, we propose a novel mechanism to achieve the association between sensor data and domain ontology. In our approach, we first classify the sensor data by making them as SSN (Semantic Sensor Network) ontology instances, and map the corresponding instances to the concepts in the domain ontology. Secondly, a multi-strategy similarity calculation method is used to evaluate the similarity of the concept pairs between the domain ontologies at multiple levels. Finally, the set of concept pairs with a high similarity is selected by the analytic hierarchy process to construct the mapping relationship between the domain ontologies, and then the correlation between sensor data and domain ontologies are constructed. Using the method presented in this paper, we perform sensor data correlation experiments with a simulator for a real world scenario. By comparison to other methods, the experimental results confirm the effectiveness of the proposed approach.

IARNN-Based Semantic-Containing Double-Level Embedding Bi-LSTM for Question-and-Answer Matching.

We propose a novel end-to-end approach, namely, the semantic-containing double-level embedding Bi-LSTM model (SCDE-Bi-LSTM), to solve the three key problems of Q&A matching in the Chinese medical field. In the similarity calculation of the Q&A core module, we propose a text similarity calculation method that contains semantic information, to solve the problem that previous Q&A methods do not incorporate the deep information of a sentence into the similarity calculations. For the sentence vector representation module, we present a double-level embedding sentence representation method to reduce the error caused by Chinese medical word segmentation. In addition, due to the problem of the attention mechanism tending to cause backward deviation of the features, we propose an improved algorithm based on Bi-LSTM in the feature extraction stage. The Q&A framework proposed in this paper not only retains important timing features but also loses low-frequency features and noise. Additionally, it is applicable to different domains. To verify the framework, extensive Chinese medical Q&A corpora are created. We run several state-of-the-art Q&A methods as contrastive experiments on the medical corpora and the current popular insuranceQA dataset under different performance measures. The experimental results on the medical corpora show that our framework significantly outperforms several strong baselines and achieves an improvement of top-1 accuracy of up to 14%, reaching 79.15%.

An improved similarity calculation method for collaborative filtering- based recommendation, considering neighbor’s liking and disliking of categorical attributes of items

Similarity measures play an important role in the accuracy of collaborative filtering based recommendation. Due to non-availability of adequate co-rated users, the accuracy of collaborative filtering decreases because it introduces the sparsity problem. In certain cases, if there is no similar user, recommendation is not possible. In order to abate the issue, we propose a novel approach that calculates the similarity between users not only based on the items rather the attributes of the items. To calculate the similarity more accurately, users’ liking and disliking of the similar attributes of a particular item are considered separately. This approach is particularly useful when there are no co-rated users in the similarity dataset. The performance of the proposed algorithm is tested on the MovieLens dataset using two accuracy metrics, MAE and RMSE.

A incremental clustering method for preference similar users

With the development of social networks, services and applications based on social networks have become more and more abundant. While providing convenient services and applications for users, the providers have also accumulated many user information. How to use information to provide users with accurate and high-quality services is an issue that needs to be solved. The division of user groups is the basis for solving this problem. User preference is the important information for division, and clustering is the means of division. In order to solve the problem of user group partition efficiency in big data environment, this paper presents a new user preference similarity calculation method. On this basis, initial clustering based on Fast Unfolding algorithm and how to determine the attribution classes of update nodes based on the initial clustering results, the clustering method of incrementally preferring similar users is proposed. Experimental results show that the proposed algorithm greatly improves the division efficiency of user groups.

Quantitative similarity calculation method for trajectory-directed line using sketch retrieval

Moving objects may have various kinds of topological events when they move relative to a directed line, such as moving in/out, folding back and stopping. This research develops a trajectory-directed line query mode using sketch-based retrieval in order to realize the query of topological relationship between a trajectory and a directed line. The research also proposes a quantitative similarity calculation method for trajectory-directed line. The sketched stroke-directed line is used as input query condition and is described as a sequence of key points for semantic association. The trajectory-directed line to be queried is also described as a sequence of key points. The time sequence distance metric is used to calculate the similarity between them. For the trajectory-directed line with different number of key points, the similarity calculation is carried out using two kinds of distance matching methods. The results show that the method is simple and feasible, and can realize the quantitative query of trajectory data.

Collaborative Filtering Based on Gaussian Mixture Model and Improved Jaccard Similarity

The recommender systems play an important role in our lives, since it can quickly help users find what they are interested in. Collaborative filtering has become one of the most widely used algorithms in recommender systems due to its simplicity and efficiency. However, when the user's rating data is sparse, the accuracy of the collaborative filtering algorithm for predictive rating is badly reduced. In addition, the similarity calculation method is another important factor that affects the accuracy of the collaborative filtering algorithm recommendation. Faced with these problems, we propose a new collaborative filtering algorithm which based on Gaussian mixture model and improved Jaccard similarity. The proposed model uses Gaussian mixture model to cluster users and items respectively and extracts new features to build a new interaction matrix, which effectively solves the impact of rating data sparsity on collaborative filtering algorithms. Meanwhile, a new similarity calculation method is proposed, which is combined by triangle similarity and Jaccard similarity. Compare our proposed model with four models based on collaborative filtering algorithms on three public datasets. The experimental results show that the proposed model not only mitigates the sparseness of the data, but also improves the accuracy of the rating prediction.