-
Notifications
You must be signed in to change notification settings - Fork 4
Frontend Research Paper Section
The Lucify front end was developed to showcase the power of LUCAS in an informational and interactive way. Lucify is connected to Yelp to get reviews of real life businesses which we analyze through LUCAS to display a variety of information to the user. On the main page of Lucify a user is able to search yelps database of business or paste a link to a business as well as select the model that LUCAS will use for analysis. After a user has selected a business the business information will be sent to LUCAS for analysis and based on the returned results Lucify will generate a results page for the business.
On the results page there are a few main parts. The information about the business is available at the top as well as a modified rating of the business gathered from only the genuine reviews. The main feature of Lucify are the chart which provide an at glance overview of the business. The first is chart is a overview Dot chart that show the review distribution of LUCAS's confidence in the review be it genuine or deceptive. The second chart is a distribution of review and business rating across time that allows the user to spot anomalies in either rating or review frequency. Through these charts a user can filter down to only the specific reviews they want to see and analyze in the next view.
All reviews are displayed in order from most genuine to most deceptive and through filtering a user can choose to only see the reviews they are interested in. Every review shows information about the user who wrote the review including badges derived from doing statistical calculation on the users overall activity on Yelp. These badges allow Lucify to enrich the information provided from LUCAS as well as provide a verdict on the users activity and trustfulness. The text of the review is displayed in a way where the words deemed by LUCAS as the most impactful on the deceptiveness or genuineness of the review are highlighted. After the reviews there is a Word Cloud that displays all of these impactful words in a chart to allow comparison across other reviews as well as giving a general overview of these words in the reviews.
The Lucify front end is used to showcase the power of LUCAS. Lucify gets businesses from Yelp and through LUCAS analyzes the reviews and displays the results in an engaging fashion. Lucify aims to highlight LUCAS ability to detect fake reviews through the means of colour, data visualizations and interaction. Lucify also enriches the LUCAS data by performing statistical analysis on the users who wrote the retrived reviews. This adds a layer of transparency to the data as well as allowing us to give a more definite verdict on the review itself.
- ACLSW 2019
- Our datasets
- Experiment Results
- Research Analysis
- Hypothesis
- Machine Learning
- Deep Learning
- Paper Section Drafts
- Word Embeddings
- References/Resources
- Correspondence with H. Aghakhani
- The Gotcha! Collection