PLAY STORE BOT DETECTION & AUTOMATED REPLY SYSTEM

Mobile app development is increasingly growing, as there are 3.48M apps present on the play store. Popular Applications have Millions of Reviews. So, it is difficult for a developer to reply to every review due to a lack of resources and a large number of reviews. This problem can be solved to

Project Title

PLAY STORE BOT DETECTION & AUTOMATED REPLY SYSTEM

Project Area of Specialization

Artificial Intelligence

Project Summary

Mobile app development is increasingly growing, as there are 3.48M apps present on the play store. Popular Applications have Millions of Reviews. So, it is difficult for a developer to reply to every review due to a lack of resources and a large number of reviews. This problem can be solved to some extent by developing an automated reply system that can automatically generate a reply to user reviews on the play store.

By using the latest machine learning algorithms, we will train our system to reply to user reviews. Mostly Previous reply systems are trained to reviews that are only classified into positive, negative, and neutral sentiments. But the user reviews can be classified into many other categories like a bug report, suggestion, etc. so we want to build a system that can generate a reply to classified categories like a bug report, suggestion, etc.

Project Objectives

In this modern era, we can solve this problem by automated reply generation. The system can automatically generate replies to thousand or millions of reviews. The system will help the app developer automatically generate replies to user reviews that are posted on his app.

We will try to improve the weaknesses of previous systems and will generate replies using two approaches. The system will be developed using the machine-learning approach and Markov chain-based approach.

Project Implementation Method

We will use an iterative approach for the proposed system. We will use Python language for developing this system due to the support of powerful libraries. We will use python machine-learning libraries to build our machine-learning algorithm. We will use the google play scraper API for getting the developer replies from the play store.

Tools and Technologies

Table 2Tools and Technologies for Proposed Project

Tools

And

Technologies

Spyder

Jupyter Notebook

Pycharm

MS Word

MS PowerPoint

Technology

Python

Python Libraries

Django

Benefits of the Project

• The main advantage of this system is that developers do not manually have to reply to each user review on the app.
• The user will appreciate the developer's reply due to which app’s rating will be increased and the user will be satisfied.
• As we have not seen any research work done on play store bot detection. So it will be helpful for future researchers to work on bot detection in the play store.
• The developers will not have to buy the available subscriptions, as this system will be open source.
• If the reply generated by this system is not suitable for any review then the developer can modify the review.
• As the studies showed that less than 1% of reviews are responded to. With this system, we can increase this percentage.

Technical Details of Final Deliverable

Module 1: Scraper

We will scrape the developer replies from the google play store using the google play scraper API. We will build a scraper that can scrape up to 50 apps’ data at a time. If someone wants to scrape the app’s data, he will just simply puts the search in the search bar. The working of this scraper

It will search for the tag in the play store.

Module 2: Developer Replies Dataset

For training our machine-learning algorithm, we need a dataset of developer replies. No such type of dataset is available on the internet. Firstly, we will define categories for replies. Then we will make a dataset of developer replies.

Module 3: Bot Detection

Firstly, we will research how can we detect a bot used by a developer in the play store. We will define the rules for bot detection. Based on these rules, we will check whether if some developers used a bot to reply to user reviews or not.

Module 4: Markov Chain Model

We will build the Markov chain model for generating a reply for user reviews. This will be the probability based reply model.

Module 5: Machine-learning Model

The machine-learning model will reply to the user reviews. The machine-learning model will be trained using the dataset of developer replies. We will create our dataset of developer replies.

Module 6: Web App

Our whole project will be in the form of a web app that will include all these modules. There will be a separate portion for each model in this app.

Final Deliverable of the Project

Software System

Core Industry

IT

Other Industries

Core Technology

Artificial Intelligence(AI)

Other Technologies

Sustainable Development Goals

Decent Work and Economic Growth, Industry, Innovation and Infrastructure

Required Resources

If you need this project, please contact me on contact@adikhanofficial.com