No abstract

<span>Rate of employment is a strong indicator of economic stability of a country. It relates to the number of volumes of produced products and services. If the unemployment rate is high, the amount of gross domestic product (GDP) of a country may be declined. One of the main factors that contributes to low rate of employment is the mismatch between job seeker and the requirement of the job applied.  This is due to the limited analysis performed on the relevant information on job advertisement; such as, skills, responsibilities of the job, location and expectation of the employers. The obscure job descriptions provided in the advertisement may result in application of unsuitable candidates that can cause rejection of the candidate and the potential employer may take a long time to filter and evaluate the applications. A system that is able to provide relevant information in a simple and catchy way is needed to simplify the task of job searching. In this paper we proposed a text analytics technique to extract users’ comments from social media such as Twitter and Facebook on job advertisement. The result is then displayed in a geotagged map that can reveal the density of job availability based on geographical location. The job seekers can easily observe and select their desired job location. The initial system shows potential of the inclusion of the proposed approach in job advertisement websites. In comparison to other job searching websites, this system can provide additional information on public view about the advertised job obtained from the social media text analytics. With this additional information, jobseekers have more confidence in job selection and allows employers to receive more suitable candidates for the available positions. It is hoped that the proposed system can tailor the job advertisements to the need of the jobseeker and making the job application more relevant hence reducing the potential employers’ processing time.</span>

No abstract

Facebook has become a popular platform in communicating information. People can express their opinions using texts, symbols, pictures and emoticons via Facebook posts and comments. These expressions allow sentiment analysis to be performed by collecting the data to obtain the public's opinions and emotions toward certain issues. Due to a huge amount of data obtained from Facebook, proper approaches are required to cater the texts and symbols used in the comments. There are also limited amount of dictionary on Malay texts which make it more challenging to process and classify the positive and negative words used in the comments. Thus, hybrid approach is applied during the data processing to visualize the results. In this work, a combination of lexicon-based approach and Naïve Bayes are used. This study focuses on analyzing the public's sentiments on crime news in Facebook by using word cloud visualization. The visualization displays important words used in a form of a word cloud. Moreover, the percentage of positive and negative words existed in the comments is also shown as part of the visualization results.