Search

Most consumers depend on online reviews posted on e-commerce websites when determining whether or not to buy a service or a product. Moreover, due to the presence of fraudulent (deceptive) reviews, the fundamental problem in such reviews is not fully addressed. Thus, deceptive reviews present wrong and misguiding opinions that are harmful to consumers and e-commerce. People called fraudsters who intentionally write deceptive reviews to target and deceive potential consumers, as they target businesses that have a well-built reputation or fame for their personal promotion, create such reviews. Therefore, developing a deceptive review detection system is essential for identifying and classifying online product reviews as truthful or fake/deceptive reviews. The main objective of this research work is to analyze and identify online deceptive reviews in electronic product reviews in the Amazon and Yelp domains. For this purpose, two experiments were conducted individually. The first was executed on standard Yelp product reviews. The second was performed on Amazon product review datasets. For this dataset, we created and labeled it using a deceptiveness score calculated based on features extracted from the review text using the linguistic inquiry and word count (LIWC) tool. These features were authenticity, negative words, comparing words negation words, analytical thinking, and positive words as well as the given rating value by a user. The recurrent neural network, bidirectional long short-term memory (RNN-BLSTM) model, was used to both datasets in order to conduct the evaluation. The application of this model was contingent upon the learning of words embedding of the review text. Finally, we evaluated the RNN-BLSTM model’s performance using the Yelp and Amazon datasets and compared the results. The results were 89.6% regarding testing accuracy for both datasets. From our experimental results, we observed that the LIWC feature with word embedding in the review text provided better accuracy performance compared with other existing methods.

Recently social media plays a major role and providing information during disasters. This paper mainly focuses on how people used social media, especially Twitter, in response to the country's worst flood, Earthquake that had occurred recently. And these tweets collecting analyzed using machine learning algorithms such as Naïve Bayes, Random Forests, Decision Tree, sentiment Analysis .during the disaster social media provides a surplus of information which includes information about the natural disaster, affected people's emotions, and relief efforts. And collect the tweets relating to disasters and build the sentimental classifier to categorize the user’s emotions during disaster based on various distress levels. Various analysis techniques are applied in collecting tweets. Keywords: Twitter Analysis, Sentiment Analysis, Random Forests, Naive Bayes, Decision Tree.

In the recent years, the carcinoma disease in women has significantly observed and it has also caused deaths as per WHO 2018 report of carcinoma statistics recorded approximately 2 lakhs registered cases and around 90000 reported deaths. The speed of survival has become very difficult at higher stages of growth and quite 45% of women's in India suffer from stage 3 and 4 of carcinoma. The target of this research is to deliver a report on carcinoma on the basis of the performance of Support Vector Machine [SVM] methodology and random forest using 5 folds, 10 folds, 20 folds with a training set size 50, 60, 70, 80 and 90 respectively. These techniques have achieved an accuracy of 94.5% and 98.40% through the cross validation of Support Vector Machine [SVM] and Random Forest [RF] method.

Internet of Things refers to the way that more and more physical devices are collecting and exchanging data over the internet. Internet of Things will have an increasing impact on bandwidth. Many Internet of Things devices operate wirelessly, while others are connected. Most IoT devices use less bandwidth, but many devices going online mean high bandwidth will be needed. As IoT grows, it will be necessary to have a platform which can accommodate this huge change. Due to the development of technology amount of data that is transmitted by devices is increased, which will need for increased bandwidth. For example, when smartphones start transmitting images and streaming video, need for bandwidth increases tremendously. There is no particular solution available for spectrum predictions. In this paper, we propose a machine learning prediction algorithm for internet bandwidth.

Online product reviews play a major role in the success or failure of an E-commerce business. Before procuring products or services, the shoppers usually go through the online reviews posted by previous customers to get recommendations of the details of products and make purchasing decisions. Nevertheless, it is possible to enhance or hamper specific E-business products by posting fake reviews, which can be written by persons called fraudsters. These reviews can cause financial loss to E-commerce businesses and misguide consumers to take the wrong decision to search for alternative products. Thus, developing a fake review detection system is ultimately required for E-commerce business. The proposed methodology has used four standard fake review datasets of multidomains include hotels, restaurants, Yelp, and Amazon. Further, preprocessing methods such as stopword removal, punctuation removal, and tokenization have performed as well as padding sequence method for making the input sequence has fixed length during training, validation, and testing the model. As this methodology uses different sizes of datasets, various input word-embedding matrices of n-gram features of the review’s text are developed and created with help of word-embedding layer that is one component of the proposed model. Convolutional and max-pooling layers of the CNN technique are implemented for dimensionality reduction and feature extraction, respectively. Based on gate mechanisms, the LSTM layer is combined with the CNN technique for learning and handling the contextual information of n-gram features of the review’s text. Finally, a sigmoid activation function as the last layer of the proposed model receives the input sequences from the previous layer and performs binary classification task of review text into fake or truthful. In this paper, the proposed CNN-LSTM model was evaluated in two types of experiments, in-domain and cross-domain experiments. For an in-domain experiment, the model is applied on each dataset individually, while in the case of a cross-domain experiment, all datasets are gathered and put into a single data frame and evaluated entirely. The testing results of the model in-domain experiment datasets were 77%, 85%, 86%, and 87% in the terms of accuracy for restaurant, hotel, Yelp, and Amazon datasets, respectively. Concerning the cross-domain experiment, the proposed model has attained 89% accuracy. Furthermore, comparative analysis of the results of in-domain experiments with existing approaches has been done based on accuracy metric and, it is observed that the proposed model outperformed the compared methods.

This paper provides a detailed introduction, significance and research progression of record de-duplication (RDD) as well as record linkage (RL) process. The basic study starts with the experimental analysis of various Blocking and Indexing techniques for Record de-duplication process, where Sorted Neighborhood Method (SNM) is found to be the best choice among all the methods. SNM is further improved using Adaptive variants of SNM. The advancements in record de-duplication are further explored and various methods for it are reviewed and implemented. The major two contributions in the unsupervised record de-duplication, FDJ and OATF are implemented and compared where it is observed that OATF which is a completely automated and unsupervised approach performs equally well as compared to unsupervised FDJ approach, where limited automation is achieved.

In the human body, abnormal growth of any cell is called cancer. Malignant region is nothing but the location of extra growth of any cell in the body. In this paper we focus on the different types of malignant region detection using thermal infrared images. We analyze different imaging techniques like thermal, X-ray, magnetic resonance, optical imaging, etc. of abnormal cell in the human body. It is analyzed that malignant regions help to detect cancer cells present within the body. The early detection of malignant region helps in saving the life of an individual. To obtain thermal images, a thermal camera is able to detect the 0.1 °C lowest temperature in the human body. The skin temperature which varies due to physiological problems and varies due to emotions will be captured through thermal camera and process.

Human body has a unique defined structure which can be differentiated from any unwanted growth of different body tissues or muscles or sometimes diseased cells. In medical science, unwanted growth of mass in the body is termed as malignant region. Abnormal growth of any cell is called as cancer. This abnormal growth of the cancer cell can affect any part/location of the body. This different location is called as malignant region. In this paper we did study on different imaging techniques focus on the different types of malignant region discovery methods such as thermal imaging, X-Ray imaging, Magnetic Resonance Imaging (MRI), optical imaging. Also, how the thermal image detects the abnormal cell in the human body. It is analyzed that malignant regions of the body help to prove the cancer cells those are present within the body. The early detection of malignant region helps in saving the life of an individual. To obtain thermal images, a thermal camera is used. Capacity of thermal camera is able to detect wide range of temperature of our body from lowest 0.1 °C to highest temperature. The temperature can vary due to physiological and emotional state in the human body.

A heart disease called cardiovascular diseases (CVD) is another leading cause for the death. There are several reasons that lead the CVD in human beings. The early detect of CVD helps to take necessary medical attentions to prevent the harms. The conventional techniques for CVD detection were manual and expensive which often delivers the inaccurate diagnosis. Since from the last decade the other inexpensive Computer Aided Diagnosis (CAD) based methods gained significant medical attentions. The CAD based techniques mainly based on raw Electro Cardiogram (ECG) signals of patient for the accurate and economical detection of CVD at early stage. In recent past, there are several CAD systems designed for CVD diagnosis utilizing raw ECG signals, however accuracy of CVD detection utilizing ECG bothered through several issues of research like QRS beats extraction, artefacts, efficient features extraction. This research paper present CVD novel framework, utilizing raw ECG signals and designed hybrid pre-processing algorithm for extracting artefacts and noise through raw ECG signal. Further designed simple and efficient dynamic thresholding based technique to extract the beats such as Q, R, S, and ST segment through pre-processing ECG signal. Third step perform the fusion of extracted beats and apply the feature extraction method called Normalized Higher Order Statistic (NHOS). The normalized HOS techniques asses the complexity among all the QRS based beats and delivers the more unique features for the accuracy enhancement. The final step is the classification by using five different classifiers for the CVD detection. The simulation results presented in this paper demonstrate that proposed framework achieved the significant accuracy improvement.

Presence of missing values in the dataset leads to difficult for data analysis in data mining task. In this research work, student dataset is taken contains marks of four different subjects in engineering college. Mean, Mode, Median Imputation were used to deal with challenges of incomplete data. By using MSE and RMSE on dataset using with proposed Method and imputation methods like Mean, Mode, and Median Imputation on the dataset and found out to be values of Mean Squared Error and Root Mean Squared Error for the dataset. Accuracy also found out to be using Proposed Method with Imputation Technique. Experimental observation it was found that, MSE and RMSE gradually decreases when size of the databases is gradually increases by using proposed Method. Also MSE and RMSE gradually increase when size of the databases is gradually increases by using simple imputation technique. Accuracy is also increases with increases size of the databases.

Information retrieval and forecasting in real time is becoming the fastest and most efficient way to obtain useful knowledge of what is happening now, allowing organizations to react quickly when problem appears which help to improve their performance. There is enormous amount of data in the form of tweets. It builds data processing system that creates informative data about the cricket test matches. Using twitter data, the authors find the sentiments or polarity of fans posting tweets related to game. Polarity is given as positive, negative and neutral. The authors also analyze the feelings or emotions of people posting tweets. Emotions are given as anger, disgust, fear, joy, sadness, surprise and unknown. Machine learning algorithm (Bayes) using R technology shows the accuracy when trained with emotion data.

One of the leading causes of death worldwide is different types of heart diseases. Such diseases are called cardiovascular diseases (CVD). Thus, the accurate diagnosis of CVD is important at the early stages to prevent from any harm. The traditional methods for CVD diagnosis are inaccurate and expensive. The Electro Cardiogram (ECG) is an inexpensive way for the CVD diagnosis. The ECG data is effectively used with the Computer Aided Diagnosis (CAD) systems for the accurate and early prediction of CVD. ECG composed of important heart-related beats which can assist in evaluating the behavior of heart. In the recent past, there are several CAD systems designed for CVD diagnosis using the raw ECG signals, and still, the number of research works going on. The CAD system for CVD analysis is composed of three main steps pre-processing, future extractions, and classification. The pre-processing method helps to improve the chances of accurate prediction, as the presence of irrelevant raw data in the original signal may lead to inaccurate outcomes. The outcome of this paper is a practical implementation and evaluation of hybrid filtering method designed for ECG signal denoising.

Sentiment lexicon is referred as Lexical Resource for sentiment analysis. It contains lexical units with their sentiment orientations. This paper introduces different sentiment lexicons with their explanations and state of art in the field of sentiment analysis and opinion mining. In Natural Language processing different kinds of lexicons plays important role in extracting useful information from text data. Lexicons are used in the pre-processing task which is a very important in text mining.

Sentiment analysis is a widely used phenomenon for analyzing online user responses to infer collective response and it is used in various applications. Negation is a very common morphological creation that affects polarity. This research paper focuses on sentence level negation identification from news articles this work uses online news articles Data from BBC news. Results are analyzed using Machine Learning Algorithms like Support vector Machine and Naive Bayes. Support Vector Machine achieves 96.46% accuracy and Naive Bayes achieves 94.16%.

Sentiment Analysis and Opinion Mining is a most popular field to analyze and find out insights from text data from various sources like Facebook, Twitter, and Amazon, etc. It plays a vital role in enabling the businesses to work actively on improving the business strategy and gain an in-depth insight of the buyer’s feedback about their product. It involves computational study of behavior of an individual in terms of his buying interest and then mining his opinions about a company’s business entity. This entity can be visualized as an event, individual, blog post or product experience. In this paper, Dataset has taken from Amazon which contains reviews of Camera, Laptops, Mobile phones, tablets, TVs, video surveillance. After preprocessing we applied machine learning algorithms to classify reviews that are positive or negative. This paper concludes that, Machine Learning Techniques gives best results to classify the Products Reviews. Naive Bayes got accuracy 98.17% and Support Vector machine got accuracy 93.54% for Camera Reviews.

Recommender systems are the software engines and approaches for providing suggestion of products to the user which might be most probably matched with the user’s like/choice. Generally, Recommendation system gives suggestions on content like what items to buy, which music to listen or what online news to read. Depending on the users liking the items are suggested. In this paper we will be discussing about the research done in the Recommendation System its approaches and the techniques used. At the end we will go through the main challenges and the limitation of the system.

User generated media like blogs, discussion forums is used as a tool by people to communicate their experiences with others. Presence of such huge data on Internet demands proper means to generate processed information from the unstructured data. What users need is more than mere sentiments. They need to know public opinion or experience of various aspects of a product like how is camera quality of the phone or energy efficiency of electronic products. For meeting the high demands of users, various techniques have been proposed till date. In this paper we are evaluating, all these techniques that discover topic along with sentiment associated with it. Many models were proposed to incorporate sentiment analysis with topic model to find aspects of a product and users sentiment about the aspect. Results of these models can be beneficial for various industries as well as users.

Now a days, huge amount data has generated on the internet and it is important to extract useful information from that huge data. Different data mining techniques are used to extract and implement to solve divers types of problems. In the era of News and blogs, there is need to extract news and need to analyze to determine opinion of that news reviews. Sentiment analysis finds an opinion i.e. positive or negative about particular subject. Negation is a very common morphological creation that affects polarity and therefore, needs to be taken into reflection in sentiment analysis. Automatic detection of negation from News article is a need for different types of text processing applications including Sentiment Analysis and Opinion Mining. Our system uses online news databases from one resources namely BBC news. While handling news articles, we executed three subtasks namely categorizing the objective, separation of good and bad news content from the articles and performed preprocessing of data is cleaned to get only what is required for analysis, Steps like tokenization, stop word removal etc. The currently work focuses on different computational methods modeling negation in sentiment analysis. Especially, on aspects level of representation used for sentiment analysis, negation word recognition and scope of negation and identification.

contain huge amount of information on Web sites the user can retrieve this with help of the search input query to Web databases & fetch the relevant information. Perhaps Web databases return the multiple search output records dynamically on Web browser, these search record are containing the Deep Web pages in the form of HTML pages. It is time consuming &human efforts are involved. The traditional search engine does not index the hidden Web pages from Web databases, such as (Google, Yahoo etc.). Many existing proposed techniques have addressed the problem of how to extract efficient structure data from Deep Web. The deep web refers to the hidden database used by web sites. But the information extraction & annotation is key challenge in web mining. The information retrieval should be done automatically & arrange in a systematic way for further processing. Various methodologies like wrapper induction is been induced. The labeling is done to the extracted information as per the concept.Various types of annotators are used on the basis of the data to be annotated. In this paper survey the automatic annotation approach on the basis of different feature of text node and data units.

Anomaly detection is a crucial analysis topic in the field of data mining as well as machine learning. Several real-world applications like Intrusion or MasterCard fraud detection need a good and efficient framework to spot deviated data instances. A good anomaly detection methodology must be able to accurately establish many varieties of anomalies, robust, need comparatively very little resources, and perform detection in period of time. In this paper we proposed the idea of combining the two different algorithms i.e. Median Based Outlier Detection and Online Oversampling PCA for effective detection of anomaly in online updating mode. Median Based outlier detection uses the interquartile range which is a measure of statistical dispersion being equal to the difference between the upper and lower quartiles. Whereas oversampling PCA does not need to store the entire covariance matrix or data matrix and thus this approach is a more useful in online or large scale problem. Compared with other anomaly detection algorithm our experimental result verifies the feasibility of our proposed method.

The live news data is vital role in the movement of stock prices. The real time unstructured data is generated through electronic and online news sources. Text mining is used for preprocessing of news stories from web sources. The visualization of stock trends can be correlated with actual market prices. The proposed analysis on current news stories helps to predict stock trends. Other techniques like tagging of stock related terms can be added for improvement in results. Stock market trends can be captured with help of this technique.

Record de-duplication is a process of identification and removal of duplicates from the given dataset in a data warehouse environment. The term record linkage is also used in the same context, the difference between record de-duplication and record linkage is that the former is used when the duplicates are to be removed from one dataset while the later is used when the duplicates are to be removed from several different datasets that refer to the same entity. Both the processes de-duplication and record linkage are important during data profiling stage of a data warehouse and assure the quality of data without repetition which in turn leads to better decision making. Record de-duplication is focused for the presented research. The Efficiency of Record de-duplication is based on several criteria such as number of comparisons needed, time and cost of comparison, accuracy level of de- duplication, time and space complexity for identification of true duplicates. In this paper we have explored the several indexing techniques which are intended to make less number of comparisons to identify duplicates from the given dataset. Peter Christen has surveyed and experimented six different indexing techniques [1] such as Sorted Neighborhood indexing, Suffix Array indexing, Q Gram based indexing, Canopy Clustering, Threshold based indexing, and String Map based indexing. In this paper, we have studied and implemented Sorted Neighborhood based de-duplication techniques in detail. During this implementation Adaptive and Non-Adaptive Sorted Neighborhood Methods are experimented and validated. Accumulative Adaptive SNM (AASNM), Incrementally Adaptive SNM (IASNM)[16] are adaptive versions of SNM while Duplicate Count Strategy (DCS) [4] is a Non Adaptive SNM. A Group based Accumulative Adaptive Method (GAASNM) is proposed to minimize the record comparisons.

Stock data analysis is challenging research area. The proposed analysis works on online news data for prediction of stock market states such as high, low etc. The Hidden Markov Model along with features extracted such as TF-IDF is used to find out next day's stock market value for group of companies. The method can be further extended to adjustment of probability values to calculate tuned model for prediction.

The various machine learning algorithms are used for prediction of stock market movement. The data collected for stock market is in the form of breaking news from various finance websites. The TF-IDF features extracted from online news data are used for creation of HMM model along with log likelihood values. The next day's stock price is predicted as either higher or lower than current day's stock price. Results obtained from proposed model is compared with results from other machine learning predictive techniques such as random forest, KNN, multiple regression, bagging and boosting. The proposed model produces approximately 70% of accurate prediction. The captured features are graphically represented with word cloud. The results can be further improved with the use of deep learning ensemble methods.

Now a day's Web technology is getting an emergence importance in day to day life! Everyone is familiar with surfing the Web, uploading personal or important data on the Web, sharing data with friends on social communities. Indian online news Web papers are producing more data every day on the Web. There are various technologies & researches which are focusing on the extraction of relevant information from large web data storage. But still there is requirement of availability of automatic annotation of this extracted information into a systematic way so to be processed further for various purposes. This paper provides an effective approach for the Indian online newspapers which extract contents from news web databases. First, we browse Web pages as per the input URL given by user. Next, we generate a DOM tree of the news Web page data. And at last, we not only identify and extract valuable news from the Indian news web pages but also remove noisy data. Moreover, in this paper we proposed the novel approach for extract data from online Indian newspapers written in the many popular languages such as Marathi, Hindi, Tamil, Gujarati, Kannada, Oriya, Telugu, Punjabi, etc. Experimental results can be analysed much easily on this domain. This proposed system is very first attempt in an India for news extraction from online web pages available in various Indian language.

For numerical record fields such as date and age, many types of error are likely to yield small numerical differences between observed and true values. If, for example, two different sources provide separate case reports related to the same incident, the dates of onset may not match perfectly but are more likely to differ by a few days than by several years. In order to tackle the variations in numbers a few methods are available. The paper proposes a new normalization technique useful for the numerical record. A Comparison of Distance with the Smith Waterman Distance shows significant increase in the weight by the present technique.

Background : Brain computer interfacing is a system that acquires and analyzes neural signals to create a communication channel directly between the brain and the computer. The EEG records the electrical fields generated by the nerve cells. With the help of Fourier Transformation the EEG signals are classified into four different frequency bands. Purpose: The main purpose of the present paper is to report results related to classification of EEG signals of different people subjected to different conditions. Methods: The experiment has been done on 10 subjects having activities related to hearing music chosen from categories of patriotic, happy,romantic and sad songs along with relaxation activity. 19 electrodes have been used under (10-20) International Standard. The δ , θ , α and β components of EEG signals to these activities have been determined. Different statistical methods including linear discriminate analysis have been tested for classification. Results: Result of the Linear Discriminant Analysis (LDA) made four groups of all modes (Relaxation, Happy, Sad, Patriotic and Romantic Song) labeled group1, Group2, Group3 and Group4 of all ten electrodes for Delta, Theta, alpha and Beta frequencies. Conclusion: The study may be used for the development of activities induced mood recognition (AIMR) system from the EEG signal. doi : 10.5214/ans.0972.7531.1017206 Competing interests: None. Source of Funding: None Received Date: 13 March 2010 Revised Date: 20 June 2010 Accepted Date: 10 July 2010