Idea mining is a new and interesting field in the areas of information retrieval research. The thoughts of people are helpful to improve strategic decision making. This paper demonstrates the efficient computational methods of idea characterization based concept by extracting the interesting hidden data from unstructured texts which come in many forms and sizes. It may be stored in patents, publications, reports, documents, Internet etc. We briefly discussed a number of successful text mining tools and text classification to extract the idea with a combination of idea mining measures.

Recently, many researchers have shown interest in using lexical dictionary for sentiment analysis. The SentiWordNet is the most used sentiment lexical to determine the polarity of texts. However, there are huge number of terms in the corpus vocabulary that are not in the SentiWordNet due to the curse of dimensionality, which will limit the performance of the sentiment analysis. This paper proposed a method to enlarge the size of opinion words by learning the polarity of those non-opinion words in the vocabulary based on the SentiWordNet. The effectiveness of the method is evaluated by using the Internet Movie Review Dataset. The result is promising, showing that the proposed Senti2Vec method can be more effective than the SentiWordNet as the sentiment lexical resource.

No abstract

The world is currently progressing towards a new connectivity era where billions of sensors are connected over a network called the Internet of Things (IoT). IoT enables a wide range of physical objects and devices to be connected and monitored with insufficient spatial and temporal detail. Despite their potential to improve multiple application domains, anomalies in the devices' behaviors pose a significant challenge, especially in the smart city's domain. Many research works have been devoted to determining such anomalous behaviors; however, there is a lack of comprehensive review focusing on anomaly detection techniques using statistical and machine learning methods in the smart cities domain. This work aims to fill this gap by presenting a review of anomaly detection techniques using statistical and machine learning methods. This paper explains the essential contexts related to IoT, followed by a review of the IoT anomaly detection techniques and their challenges, types, and detection modes. The paper then presents a summary of the related works related to smart cities. Finally, the open challenges and future directions were highlighted.

The number of users of an on-line shopping websites is continuously increasing. Such website often provides facility for the users to give comments and ratings to the products being sold on the websites. This information can be useful as the recommendation for other users in making their purchase decision. This paper investigates the problem of predicting rating based on users' comments. A classifier based on information retrieval model is proposed for the prediction. In addition, the effect of integrating sentiment analysis for the rating prediction is also investigated. Based on the results, an improvement in prediction performance can be expected with sentiment analysis where an increase of 54% is achieved.