“News & Moods” predicts Labor to win both NSW state and federal elections from online news sentiment computing
Announcement posted by Lexxe Pty Ltd 14 Mar 2019
“News & Moods” predicts Labor to win both NSW state and federal elections from online news sentiment computing
Sydney, Thursday, 14 March 2019
Lexxe, a Sydney-based start-up based in Macquarie University Incubator, recently launched its news reading website News & Moods (https://www.newsandmoods.com/), which analyses news in terms of public sentiment. Crawling over 30,000 major news articles from 600 sources around the world everyday, the service can reveal prevailing sentiment – positive or negative – about popular topics being covered and discussed online.
Comparing sentiments for keyword searches over the last 7 days on Thursday morning (March 14, 2019) covering NSW Labor, NSW Liberal, Michael Daley (State Labor leader) and Gladys Berejiklian (Premier and State Liberal leader), the search engine is currently showing a win for NSW Labor in the upcoming NSW State Election on March 23rd, 2019. NSW Labor scored an aggregated sentiment score 0.07 in positive, while NSW Liberal instead received -0.01] (See diagram 1).
Comparing sentiments for keyword searches over the last 7 days also on Thursday morning (March 14, 2019) covering Labor, Liberal, Bill Shorten (Federal Labor leader) and Scott Morrison (Prime Minister and Federal Liberal leader), the platform shows a comfortable win with an aggregated sentiment score of 0.15, a positive sentiment indicated for Labor vs negative for the Coalition with an aggregated sentiment score of -0.03] (See diagram 2).
Lexxe’s founder, Dr. Hong Qiao, says “using online sentiment search as an indicator for success could be a more exact science than polling due to the huge sample sets being used to form its results. The process can provide results in realtime and can show rapidly using trends in time. We believe it is a powerful tool in predicting outcomes where public sentiment is a key influencer for results.”
Search Results explained
The predictions are made with 7 days of data prior to the clicking. So, everyday the results may vary according the sentiment computed from all the news about the two major political parties and their leaders. The closer to the date of election, the more relevant it is to the prediction of the election. The way to examine a party’s sentiment scores is by adding up the party’s score and its leader’s. Then we compare the aggregated score of the two parties.
NSW Election
Clicking on the following link, you will get a result like Diagram 1 below. But please be aware that the diagram you get will be based on the information of the time you click the link, which will be different from the one below.
Diagram 1. NSW State Election prediction with sentiment scores of keywords “NSW Labor”,
“NSW Liberal”, Michael Daley, Gladys Berejiklian, over the past 7 days
from Thursday morning (March 14, 2019).
In NSW, the figures indicated the NSW Liberals will win. You can compare NSW Labor on 0.04 + Michael Daley on 0.03 delivering an aggregated score of 0.07. Likewise, NSW Liberal on -0.01 + Gladys Berejiklian on 0.00 received -0.01 on aggregation. Therefore, the system predicted Labor to win in NSW state election, with the NSW Liberals trailing 0.08.
Federal Election
Clicking on the following link, you will get a result like Diagram 2 below. Please remember, the diagram could be different, when you click the link, because news contents change over time.
Diagram 2. Federal Election prediction with sentiment scores of keywords
Labor, Liberal, Bill Shorten, Scott Morrison, over the past 7 days
From Thursday morning (March 14, 2019).
Using the same method, now let’s look at the federal situation here with the Labor Party (Labor on 0.13 + Bill Shorten on 0.02) scoring 0.15 on aggregation, while the Liberal Party was on -0.03 on aggregation, given the Liberal Party on -0.03 and Scott Morrison on 0.00. So News & Moods predicts Labor the winner of the federal election, which was way ahead of the Liberals.
About Lexxe
News & Moods, a flagship product of Lexxe, is the first open platform in the world to show sentiment information of news. Dr. Qiao, Ph.D. in Computational Linguistics from the University of Queensland and patent holder working at Lexxe, a search engine start-up based at Macquarie University Incubator to expand the technology, foresaw the importance of sentiment information in future Internet search domain and developed the platform to be used by governments and businesses to monitor their reputation. It can also be a good tool for ordinary individual users to find out information about manufacturers and products.
For more information, please contact Dr. Hong Qiao at hong.qiao@lexxe.com.