Fake Review Detection on Google and Zillow
Budget: $750 – $1,500 USD
FAKE ONLINE REVIEW DETECTOR WANTED.
The desired outcome:
A software program that will analyze and detect fake online reviews, specifically on Google and Zillow, and then summarize that information in a report that flags suspected fake reviews.
I understand that algorithms can be trained to detect fake reviews through textual features such as excessive punctuation use, poor grammar, and an overly negative or positive tone. Algorithms can also monitor the behavioral pattern of reviewers, such as the user’s total number of reviews, review dates, and user profile details. I just don't have the expertise to create or train those algorithms.
Generative AI tools can also help in fake review detection. The algorithm first needs examples of fake reviews along with explanations focusing on the cues that might indicate if the reviews are fake. Then you can present a set of reviews and ask the algorithm to identify which ones might be fake.
This is my idea of the perfect fake online review detector program:
It will analyze profile and content data from online Google or Zillow reviews and produce an easy-to-read report.
THE PROFILE DATA will include the behavioral data of the reviewers – how many businesses they have reviewed, the types of businesses they have reviewed, and the types of reviews they leave.
When looking at all the businesses a reviewer has reviewed, the program will calculate the distances between all of those businesses. This will help to understand a group of reviewers of a particular business.
It will also analyze reviewers to see if there is a large group of reviewers who have reviewed the same businesses.
It will also show whether their profile or other reviews are blocked from public view.
Further, it will check reviewer images to see if reviewers have stolen photos from celebrities on instagram or facebook or elsewhere on the internet.
THE CONTENT DATA is found in the language of the review itself.
Machine learning can identify unique authorship styles and can examine reviews written by the same author.
Since fake reviews are generally higher in sentiment than real reviews, the program can look at sentiment to make a prediction whether a review is real or fake. It can also look at the use of exclamation marks and other indications that show over the top positive sentiment.
Lastly, the program should measure keyword density and look for keyword stuffing to see if the business is stuffing keywords into the reviews in order to rank better for specific industry-related keywords.
If it is helpful to get you started:
Kaggle has various datasets of reviews with clear licensing terms allowing commercial use.
SentiOne also specializes in collecting and analyzing online conversations, including reviews, social media posts, and other user-generated content. It also provides data for commercial use.
Thank you for your consideration and I look forward to hearing from you.
The desired outcome:
A software program that will analyze and detect fake online reviews, specifically on Google and Zillow, and then summarize that information in a report that flags suspected fake reviews.
I understand that algorithms can be trained to detect fake reviews through textual features such as excessive punctuation use, poor grammar, and an overly negative or positive tone. Algorithms can also monitor the behavioral pattern of reviewers, such as the user’s total number of reviews, review dates, and user profile details. I just don't have the expertise to create or train those algorithms.
Generative AI tools can also help in fake review detection. The algorithm first needs examples of fake reviews along with explanations focusing on the cues that might indicate if the reviews are fake. Then you can present a set of reviews and ask the algorithm to identify which ones might be fake.
This is my idea of the perfect fake online review detector program:
It will analyze profile and content data from online Google or Zillow reviews and produce an easy-to-read report.
THE PROFILE DATA will include the behavioral data of the reviewers – how many businesses they have reviewed, the types of businesses they have reviewed, and the types of reviews they leave.
When looking at all the businesses a reviewer has reviewed, the program will calculate the distances between all of those businesses. This will help to understand a group of reviewers of a particular business.
It will also analyze reviewers to see if there is a large group of reviewers who have reviewed the same businesses.
It will also show whether their profile or other reviews are blocked from public view.
Further, it will check reviewer images to see if reviewers have stolen photos from celebrities on instagram or facebook or elsewhere on the internet.
THE CONTENT DATA is found in the language of the review itself.
Machine learning can identify unique authorship styles and can examine reviews written by the same author.
Since fake reviews are generally higher in sentiment than real reviews, the program can look at sentiment to make a prediction whether a review is real or fake. It can also look at the use of exclamation marks and other indications that show over the top positive sentiment.
Lastly, the program should measure keyword density and look for keyword stuffing to see if the business is stuffing keywords into the reviews in order to rank better for specific industry-related keywords.
If it is helpful to get you started:
Kaggle has various datasets of reviews with clear licensing terms allowing commercial use.
SentiOne also specializes in collecting and analyzing online conversations, including reviews, social media posts, and other user-generated content. It also provides data for commercial use.
Thank you for your consideration and I look forward to hearing from you.