查看文章

The paper investigates the employment sentiment score and unemployment rate in Malaysia. Employment sentiment score is collected from multiple channel social media and lexicon-based technique is used to classify the sentiments into three different categories based on the keywords. Besides, unemployment rate also collected from DATASTREAM to examine the changes of unemployment rate in Malaysia. A total of 778,676 media messages are collected from January 2015 until December 2016 and the comments are transforms into sentiment score using machine learning techniques. The sentiment score of employment is analysis and compare with unemployment rate in Malaysia. Results found that are negative sentiment score on employment although the unemployment rate is less volatile in recent years.