The sale of commuter bus tickets by human ticket agents during peak periods can be stressful to the participants involved. The ticket agents need to select departure locations from a list of possible departures provided by the bus service. For each departure, there is then a list of possible destinations. The commuter bus service described in this research used Android devices with a large format screen along with a connected secure ticket printer. Many of the ticket vending stations were in rural areas catering for commuters who lived in rural areas and worked in urban areas. As such, the connectivity was often slow. The devices themselves were also constrained with many of them still running on Android Jelly Bean. This research looks at ways to increase the speed of individual ticket sales by using data mining and machine learning models on centralized servers, then publishing these models as small text files on REST servers. These small text files could then be downloaded by the Android devices each day at the start of business to assist in predicting departures and destinations to the ticket agents thereby decreasing the amount of time a customer and ticket agent needed to conduct a sale.

Did you like this research project?

Medical practice is performed based on clinical research, this being commonly based on evaluation of drugs and therapeutic procedures' effect. With the increase of technology and computational storage facilities, usage of Real World Data (RWD) and Big Data Mining (BDM) techniques is proving to be a useful tool for automated data analysis. Entities dealing daily with medical practice such as clinics and hospitals possess databases with a wellspring of information worthwhile being studied to aid clinicians establishing disease patterns identification, future trends, and therapeutic relationships. Aiming at assessing cardiovascular disease (CVD) progression of diabetic patients, a nearly 20 years old private clinic data-base was studied. Primarily goal, and subject of this paper, was the evaluation of the data-base reliability to continue the study of CVD progression. Manual inspection of the database content revealed missing and misleading fields, inconsistency of inputted instrumental data, temporal and user dependency of fields filling, particularly concerning CV data. But, statistics computed on the 20222 diabetic patients' records whose empty entries were eliminated revealed RWD conclusions coherent with published ones.

Did you like this research project?

In this paper, we introduce three maritime Port Congestion Indicators (PCIs) mined using Automatic Identification System (AIS) static and dynamic messages. The proposed indicators are spatial complexity, spatial density, and time criticality. To calculate the PCIs, we proposed three Big AIS Data mining algorithms to find the geohash area for certain precision, the convex hull area, and the average vessels proximity within the Port Area of Interest (AOI) and in the Period of Interest (POI). The indicators are calculated for the year of 2015 for three ports (Halifax, Hong Kong, and Singapore). The proposed PCIs capture the spatial complexity, spatial density, and time of service criticality. These indicators can be used by port authorities and other maritime stakeholders to alert for congestion levels that can be correlated to weather, high demand, or a sudden collapse in capacity due to strike, sabotage, or other disruptive events. We clustered the indicators for each port into three colour-coded (Green, Yellow, and Red) clusters corresponding to low, medium and high congestion levels. The centroids of these clusters can be used to predict future congestion levels of the port under consideration. To the best of our knowledge in published literature, this work is the first to introduce the application of AIS Big Data analytics to evaluate maritime port congestion levels.

Did you like this research project?

The proliferation of social networks allowed creating a big quantity of data about users and their relationships. Such data contains much private information. Therefore, anonymization is required before publishing the data for data mining purposes (scientific research, marketing, decision support etc).Most of anonymization works focus on the privacy preserving techniques that allow publishing one instance of the social network without revealing sensitive information. However to analyze the evolution of the social network sequential releases are needed. In this paper we study the problem of privacy in sequential releases of social networks which are represented as labeled bipartite graphs to model the affiliation relationship between users and interests. We propose a solution that allows publishing sequential releases of the same social network while preserving the privacy of data. We consider a set of complex queries to study the utility given by our solution. The experiments demonstrate that the utility of data is preserved as the queries can be answered with reasonable accuracy over the anonymized data.

The information publication on the Internet increases the information accessibility. However, the amount of data on the Internet increases rapidly and it becomes difficult to find the needed information among multiple possible sources. Therefore, automated tools for web mining gain their popularity and seek to simplify the gathering of the information, published on the internet. The purpose of information gathering as well as web page development technologies, format of published data and other properties might be different therefore different web mining methods exist. However, there is no clear classification or a complete list of web mining techniques therefore it is difficult to fully understand the web mining area. In this paper we present a new web mining taxonomy, designed to clarify the web mining area. The taxonomy will serve as classification methodology and will allow classification of existing web mining technologies or methods according to clearly defined criteria and categories.

Did you like this research project?

Microblogs sentiment analysis of people's attitudes, appraisals and emotions has become one of the most active research areas for business marketing, decision making, political campaigns, and alike. As people publish short snippets of texts through the social networks expressing their ideas, thoughts and opinions, an instant and reliable mining machine should be utilized. In this paper, we proposed an IoT mining machine for Twitter sentiment analysis. Firstly, we used Twitter's API for harvesting tweets in real time. Then, a mining engine was developed on the Raspberry Pi single-board microcomputer as an IoT platform due to its availability and connectivity. The IoT device was programmed for sentiment analysis and opinion mining using state-of-the-art Naïve Bayes classifier which after training was used to classify the trending tweets into either positive or negative. We used a gold standard dataset from SemEval 2017 for training our classifier which achieved 0.992 of accuracy. We aggregated the sentiments of tweets streamed in daily trend hashtags into visualized graphs. Finally, the visualized results from opinion mining were displayed on two-way smart mirror without any need for application installment. Our experimental results on the IoT mining machine demonstrate its feasibility and effectiveness.

Did you like this research project?

In this paper, a new approach for increasing the approximation accuracy with the use of computational intelligence tools is described. It is based on the compatible use of the neural-like structure of the Successive Geometric Transformations Model and the inputs polynomial extension. To implement such an extension, second degree Wiener polynomial is used. This combination improves the method accuracy for solving various tasks, such as classification and regression, including short-term and long-term prediction, dynamic pricing, as well as image recognition and image scaling, e-commerce. Due to the use of SGTM neural-like structure, the high speed of the system is maintained in both training and using modes. The simulation of the described approach is carried out on real data, the time results of the neural-like structure work and the accuracy results (MAPE, RMSE, R) are given. A comparison of the operation of the method with existing ones, such as Support vector regression, Classic linear SGTM neural-like structure, Linear regression (using Stochastic Gradient Descent), Random Forest, Multilayer Perceptron, AdaBoost are made. The advantages of the developed approach, in particular with regard to the highest accuracy among existing ones were experimentally established.

Did you like this research project?

The urgency of using e-Iearning in higher education is not in doubt at this time. The trend of using e-learning in the current practice of educational institutions and business structures of various levels has become global, due not only to the peculiarities of the development of educational systems in the world, but also to economic factors, which in many cases are the main driver of changes and innovations in the educational sphere. In spite of all the difficulties of deployment, publishing of those systems and troubles of their efficiency maintenance, their usage is spreading in tertiary education as well as in vocational education. In order to gain profits from e-Iearning materials Ukrainian universities search the possibilities to enter European e-Iearning market. For this reason it is vital to understand European e-Iearner, his needs and features. Therefore the European e-Iearner profile identification was performed on the basis of open European statistical datasets by the means of developed data conversion mechanism.The urgency of using e-Iearning in higher education is not in doubt at this time. The trend of using e-learning in the current practice of educational institutions and business structures of various levels has become global, due not only to the peculiarities of the development of educational systems in the world, but also to economic factors, which in many cases are the main driver of changes and innovations in the educational sphere. In spite of all the difficulties of deployment, publishing of those systems and troubles of their efficiency maintenance, their usage is spreading in tertiary education as well as in vocational education. In order to gain profits from e-Iearning materials Ukrainian universities search the possibilities to enter European e-Iearning market. For this reason it is vital to understand European e-Iearner, his needs and features. Therefore the European e-Iearner profile identification was performed on the basis of open European statistical datasets by the means of developed data conversion mechanism.

Did you like this research project?

Online and offline analytics have been traditionally treated separately in software architecture design, and there is no existing general architecture that can support both. Our objective is to go beyond and introduce a scalable and maintainable architecture for performing online as well as offline analysis of streaming data. In this paper, we propose a 7-layered architecture utilising microservices, publish-subscribe pattern, and persistent storage. The architecture ensures high cohesion, low coupling, and asynchronous communication between the layers, thus yielding a scalable and maintainable solution. This design can help practitioners to engage their online and offline use cases in one single architecture, and also is of interest to academics, as it is a building block for a general architecture supporting data analysis.

Did you like this research project?

The proliferation of ubiquitous Internet and mobile devices has brought about the exponential growth of individual data in big data era. The network user data has been confronted with serious privacy concerns for extracting valuable information during the process of data mining. Differential privacy preservation is a new paradigm independent of the adversaries' prior knowledge, which protects sensitive data while maintaining certain statistical properties by adding random noise. In this paper, we put forward a differential privacy preservation multiple cores DBSCAN clustering schema based on the powerful differential privacy and DBSCAN algorithm for network user data to effectively leverage the privacy leakage issue in the process of data mining, enhancing data clustering efficaciously by adding Laplace noise. We perform extensive theoretical analysis and simulations to evaluate our schema and the results show better efficiency, accuracy, and privacy preservation effect than previous schemas.

Did you like this research project?

Combining the results of different search engines in order to improve upon their performance has been the subject of many research papers. This has become known as the “Data Fusion” task, and has great promise in dealing with the vast quantity of unstructured textual data that is a feature of many Big Data scenarios. However, no universally-accepted evaluation methodology has emerged in the community. This makes it difficult to make meaningful comparisons between the various proposed techniques from reading the literature alone. Variations in the datasets, metrics, and baseline results have all contributed to this difficulty. This paper argues that a more unified approach is required, and that a centralised software platform should be developed to aid researchers in making comparisons between their algorithms and others. The desirable qualities of such a system have been identified and proposed, and an early prototype has been developed. Re-implementing algorithms published by other researchers is a great burden on those proposing new techniques. The prototype system has the potential to greatly reduce this burden and thus encourage more comparable results being generated and published more easily.

Did you like this research project?

Scholarly information usually contains millions of raw data, such as authors, papers, citations, as well as scholarly networks. With the rapid growth of the digital publishing and harvesting, how to visually present the data efficiently becomes challenging. Nowadays, various visualization techniques can be easily applied on scholarly data visualization and visual analysis, which enables scientists to have a better way to represent the structure of scholarly data sets and reveal hidden patterns in the data. In this paper, we first introduce the basic concepts and the collection of scholarly data. Then, we provide a comprehensive overview of related data visualization tools, existing techniques, as well as systems for the analyzing volumes of diverse scholarly data. Finally, open issues are discussed to pursue new solutions for abundant and complicated scholarly data visualization, as well as techniques, that support a multitude of facets.

Did you like this research project?

The universal popularity of GPS-enabled mobile devices and traffic navigation services has fueled the growth of trajectory data, as evidenced by Uber Movement and NYC taxi data release. Although trajectory data can generate valuable insights and value-added services for many, publishing this data while respecting mobile users' privacy has been a long-standing challenge. In this paper we present DP-Star, a methodical framework for publishing trajectory data with differential privacy guarantee as well as high utility preservation. DP-Star relies on a novel combination of several components. First, DP-Star's normalization algorithm uses the MDL metric to summarize raw trajectories using their representative points, thereby achieving a desirable trade-off between the preciseness and conciseness of their information content. Second, DP-Star constructs a density-aware grid which ensures spatial densities can be preserved despite the noise added to satisfy differential privacy. Third, DP-Star preserves the correlations between trajectories' end points through a private trip distribution, and intermediate points through a private Markov mobility model. Finally, DP-Star estimates users' trip lengths using a median length estimation method, and generates synthetic trajectories that preserve both differential privacy and high utility. Our experimental comparison shows that DP-Star significantly outperforms existing approaches in terms of trajectory utility and accuracy.

Did you like this research project?

In this article, a new adaptive wavelet diagnostic neuro-fuzzy network (AWDNFN) based on additive generalized neuro-models and the adaptive W-neuron is proposed. The learning algorithm of proposed AWDNFN is based on quadratic criterion. A large class of biomedical data processing tasks (in the data stream form) can be solved using this system. The AWDNFN has advantages in comparison with the existed methods, among them high-speed information processing in online mode. The experiments based on benchmark and real data are performed. The obtained results of experiments justify the advantages of proposed AWDNFN.

Did you like this research project?

As one of the fundamental tasks in text analysis, phrase mining aims at extracting quality phrases from a text corpus and has various downstream applications including information extraction/retrieval, taxonomy construction, and topic modeling. Most existing methods rely on complex, trained linguistic analyzers, and thus likely have unsatisfactory performance on text corpora of new domains and genres without extra but expensive adaption. None of the state-of-the-art models, even data-driven models, is fully automated because they require human experts for designing rules or labeling phrases. In this paper, we propose a novel framework for automated phrase mining, AutoPhrase, which supports any language as long as a general knowledge base (e.g., Wikipedia) in that language is available, while benefiting from, but not requiring, a POS tagger. Compared to the state-of-the-art methods, AutoPhrase has shown significant improvements in both effectiveness and efficiency on five real-world datasets across different domains and languages. Besides, AutoPhrase can be extended to model single-word quality phrases.

Did you like this research project?

Personalized recommendation is crucial to help users find pertinent information. It often relies on a large collection of user data, in particular users' online activity (e.g., tagging/rating/checking-in) on social media, to mine user preference. However, releasing such user activity data makes users vulnerable to inference attacks, as private data (e.g., gender) can often be inferred from the users' activity data. In this paper, we proposed PrivRank, a customizable and continuous privacy-preserving social media data publishing framework protecting users against inference attacks while enabling personalized ranking-based recommendations. Its key idea is to continuously obfuscate user activity data such that the privacy leakage of user-specified private data is minimized under a given data distortion budget, which bounds the ranking loss incurred from the data obfuscation process in order to preserve the utility of the data for enabling recommendations. An empirical evaluation on both synthetic and real-world datasets shows that our framework can efficiently provide effective and continuous protection of user-specified private data, while still preserving the utility of the obfuscated data for personalized ranking-based recommendation. Compared to state-of-the-art approaches, PrivRank achieves both a better privacy protection and a higher utility in all the ranking-based recommendation use cases we tested.

Did you like this research project?

Reviews of tourist areas are based on the author's experience; hence, when other users travel to a tourist area, they may experience it differently. Over the years, many papers analyzing tourism reviews have been published. However, there are not many papers that analyze tourism blog data and present a summary of their content. One aspect of blog content is the large quantity of data, making it difficult for users to understand the blog content in a short amount of time. Therefore, in this paper, we propose a system to analyze blogs and review data written by users based on their tourism experiences, summarize their content, and visualize the information that the blog authors intended to convey according to various tourism properties. The results of our experiments confirmed that the system proposed in this paper can reduce the time spent by an average of 91.6%.

Did you like this research project?

A technical standard is an established norm or requirement in regard to technical systems. It is usually a formal document that establishes uniform engineering or technical criteria, methods, processes and practices. Every year, more than 10K standards are released by standardization organizations, such as International Standardization Organization (ISO), American National Standards Institute (ANSI), and China National Institute of Standardization (CNIS). Thereinto, China publishes more than 4K national and industrial standards per year. Confronted with the large amount of standards, how to manage and make full use of these standards efficiently and effectively has become a major issue to be addressed. In this paper, we introduce a mining and visualizing system for large-scale Chinese technical standards from spatial and temporal perspectives. Specifically, we propose an s-index metric to measure the contribution of a drafting unit. Moreover, we develop a multiple dimensional mining and visualizing analysis system for standardization experts and users to explore regional differences and temporal trends of standards in China. Our system has run on 95K Chinese standards spanning from 2001 to 2016 and provided online service. The extensive experiments and case study demonstrate the effectiveness of our proposed system.

Did you like this research project?

Data mining techniques have been shown its success in analyzing data to assist factors and make decisions in many different applications. In this work, we analyzed universities' faculty members attitudes toward using Wikipedia in teaching in higher education. This may help in developing appropriate solutions to improve Wikipedia's effectiveness in education and provide the developers of Wikipedia with factors that may negatively impact using Wikipedia in teaching in higher education. Three different classification algorithms (Jrip Rule, Decision tree (J48) and Naïve Bayes) have been carried out in the free and open source software WEKA (Waikato Environment for Knowledge Analysis) on used dataset. The accuracy and Receiver Operating Characteristic curves for all classifiers discussed. Our results show that there are some factors influences the faculty using of Wikipedia in activity teaching which are unobserved in other studies.

Did you like this research project?

An algorithm is proposed for encrypting-decrypting images using RSA algorithm elements, as the most cryptographically resistant to unauthorized decryption related to images with strictly clear edges. It is proposed to use RSA algorithm elements as coefficients of some linear-quadratic affine transformation. The proposed algorithm has a higher cryptographic stability compared to RSA algorithm.

Did you like this research project?

In the era of big data, government, business and personal digital information will be possible for data mining. Data mining requires massive data as a support. However, the direct release of the original mass data, which usually contain some sensitive information of personal or analysis, will result in leakage of user privacy. Therefore, it is becoming increasingly important to protect privacy information in data publishing. In this paper, we focus on the multi-type self-identified format-preserving encryption. First, we introduce a multi-type self-identified format-preserving encryption system and discuss the encryption of various types of data in this system. Then, for the format preserving encryption (FPE) about Chinese name, we study from the encryption model construction and basic encryption scheme. The format-preserving encryption model about Chinese name is constructed and the concept of the name library is presented. Based on this, it is used to not only limit the message space to reduce complexity, but also ensure the cipher in accordance with the Chinese naming habits. In addition, according to the encryption and decryption model, format-preserving encryption process of Chinese name is designed. In order to add new names, the algorithm of name space expansion is proposed. Based on the Prefix, this paper put forward an algorithm named Cycle-Prefix, which enhances the security and dynamics of FPE by using two adjustment factors and the circular encryption. Compared with the traditional Prefix algorithm, experiments show that Cycle-Prefix can not only complete the task of FPE for Chinese name, but also encrypt same plain text into different ciphers under the premise of similar efficiency with Prefix.

Did you like this research project?

The Public Library of Science (PLOS) is a major publisher of academic journals with a collection of over 200,000 scientific journal articles. As the library grows, a key operational challenge for PLOS is the matching of a submitted manuscript with one of the nearly 6,000 academic editors with whom they work. We frame this problem as one of reducing distance between editors and documents in a network, where editors and documents are modeled as nodes in a graph connected by shared features. We apply Natural Language Processing (NLP) and text mining methods, such as topic modeling and word embedding, to create new nodes and edges in the network to reduce its scale, thus reducing the distance between documents and editors. These new nodes and edges form a rich semantic network that may also be used to represent the thematic content of the corpus as a whole. Our solution iterates on PLOS's current internal tool for editor matching, PLOS Match, by recommending potential editors who have edited semantically similar manuscripts in the past. Improved quality of editor matches will increase the rate at which editors accept an invitation to participate in manuscript evaluation. This will streamline PLOS's internal review process and hence speed knowledge delivery.

Did you like this research project?

Career building is the most cherished part of every engineering student. For an engineering graduate it is necessary to have immense knowledge in their domain to get placed in a reputed company. Data Mining is used to gain knowledge, find the hidden information and also this system applies data mining techniques to the academic dataset. The Academic data includes the Internal (CCET 1, CCET2 and CCET3) marks and the Assignment marks. The final semester marks are predicted from the analyzed result of each student. In order to increase the accuracy, this system introduces reweight enhanced boosting algorithm.

Did you like this research project?

Cyber attacks against sensitive data have become as serious threats all over the world due to the rising applications of computer and information technology. New malware or malicious programs are released everyday by cyber criminals through the Internet in an attempt to steal or destroy important data. Hence, research on protecting data receives tremendous interest in the cyber community. In order to cope with new variants of malicious software, machine learning techniques can be used for accurate classification and detection. This paper proposes an efficient scheme for malware detection for protecting sensitive data from malicious threats using data mining and machine learning techniques. Experimental results shows that the proposed approach gives better performance compared to other similar methods.

Did you like this research project?

Data mining techniques are used in the field of medicine for various purposes. The function of data mining in biomedical information is remarkable one. Mining association rule is one of the interesting topics. It was first proposed for market basket analysis. In this project a method has been implemented to identify risk factors having heart disease through frequent Itemset. Frequent itemsets are generated based on the attributes and minimum support value. The extracted frequent itemsets help the medical practitioner to make diagnostic decisions and determine the risk level of patients at an early stage. The proposed method can be applied to any medical dataset to predict the risk factors with risk level of the patients based on chosen factors.

Did you like this research project?

In recent time, online social networks have been affected by various unwanted threats. Although they provided us with an open platform to share our thoughts with others, however, due to misuse of this powerful resource, general users are in endangered condition. For example, YouTube has been used as a promotional ground by various artist to upload their music videos, movie trailers, etc. and viewers can post their opinion on them. Unfortunately, often malicious users use to post phishing website links, advertisements, and fraudulent information in the comments section, which may transmit viruses or malwares. So, these harmful comments need to be identified in order to continue flawless service of social media. In this study, we have been implemented several classification algorithm to sort out the spam comments on YouTube videos from the legitimate one, their performance measures have been analysed as well as performance of ensemble classifier over single classifier algorithm in the context of text classification has also been highlighted.

Did you like this research project?

The biggest concern of Network is security. Intro find the tricks and tools of the Attackers. Data Mining techniques automatically learn the pattern of the tuples and Intelligent decision are made. Supervised learning methods finds the attack based on previous knowledge and unknown attacks are detected by using Unsupervised learning. Dos, Probe and Normal data are correctly detected by maximum Data Mining algorithms, whereas True Positive Rate of R2L and U2R are very low. The paper highlight the best attribute selection method which helps to improve the accuracy of the algorithms. The Hybrid methods (Kmeans + ID3 and K-means + Support vector machine) are used to improve True Positive Rate of R2L attacks. NSL_KDD Training and Testing dataset are used.

Did you like this research project?

Data mining is one of the widespread research areas of present time as it has got wide variety of application to help people of today’s world. It is all about finding interesting hidden patterns in a huge history database. In this research work, data mining is comprehensively applicable to a domain called placement chance prediction, since taking wise career decision is so crucial for all of us for sure. A strategy to predict the overall absorption rate for every branch as well as the time it takes for all the students of a particular branch to get placed etc. are also proposed. The proposed method is tested on the data set provided by A.I.M.T college Lucknow and data is passed through the various data mining model, namely decision tree , neural network and navie bayes classifier on area of the application to a domain, development of classifier and future outcomes were also configured on this thesis. At last, this research work puts forward the data mining algorithm namely C 4.5 * stat for numeric data sets which has been proved to have competent accuracy over standard benchmarking data sets called UCI datasets. It also proposes to improve the standard C 4.5 algorithm

Did you like this research project?

Many medicinal fields comprise lots of data regarding the health care. There may be some data with which we can derive some interesting patterns which frequently go exploited. Data mining is the process of analyzing huge data sets and derive the relevant information. It becomes more prominence in case of heart disease which is the major concern behind the deaths all over the world. The term heart disease refers to various types of conditions that can affect the heart functions. This medicinal condition defines the unpredicted health conditions that directly control all the parts of the heart. It makes use of different parameters included in data mining include clustering, association rule mining and prediction which are used to predict the heart disease. Initially the heart disease data set is preprocessed to eliminate any irrelevant data. The preprocessed data is now clustered using K-means clustering algorithm. Maximal Frequent Item set Algorithm (MAFIA) is used to derive the maximal frequent patterns in data set. These results can be classified using C 4.5 training algorithm to build a decision tree. The calculated metrics proved that our designed prediction system is capable of predicting the heart disease successfully.

Did you like this research project?

To judge the performance of students in advanced education has now become a great challenge not only in academic but in curriculum activities as well. In this manner, it is important to adequately analyze the data which is utilized for educating and learning processes. Mining data from educational dataset helps to extract useful information for enhancing the teaching and learning abilities. This survey presents the overview of various Data Mining Techniques in terms of advanced education as a new research domain known as educational data mining (EDM) for foreseeing the execution behavior of the students and to find out the reasons behind their failures. The intended review of literature is assessed to briefly evaluate the work done in educational field. Through this task, we extract those factors that describe student's attainments and discover those ones who require special concern by introducing various data mining methods which assimilate classification, clustering, regression, prediction and so forth.