Now you can Instantly Chat with Ashish!
About Me
A self-driven person to achieve perfection and professionalism in the technologies. Seeking to pursue a professional career that provides me an opportunity to learn, utilize my skills, and professional growth along with the organization. Recognized b...
Show MoreSkills
Positions
Portfolio Projects
Description
To build a classification methodology to predict the type of thyroid based on the given training data. Data Analysis using python libraries is done for this project and machine learning models are compared and then on the best model hyper parameter tuning is done for prediction the type of thyroid. At the end the product as deployed on Azure cloud services.
Show More Show LessDescription
To build a regression model to predict the concrete compressive strength based on the different features in the training data. It is an AI-based web application, deployed in local system. This Web-based application has the capability to train itself with a single button click and predict results with good accuracy.
Show More Show LessDescription
To build a classification methodology to predict whether a costumer is going to be a defaulter for a payment or not. It is an AI-based web application, deployed in local system. This Web-based application has the capability to train itself with a single button click and predict results with good accuracy.
Show More Show LessDescription
To build a classification methodology to predict whether a website is a phishing website on the basis of given set of predictors. It is an AI-based web application, deployed in local system. This Web-based application has the capability to train itself with a single button click and predict results with good accuracy
Show More Show LessDescription
To build a classification methodology to predict whether a website is a phishing website on the basis of given set of predictors. It is an AI-based web application, deployed in local system. This Web-based application has the capability to train itself with a single button click and predict results with good accuracy.
Show More Show LessDescription
I tried to make a CNN model from scratch and models using Transfer learning for identification of malaria images weather they are parasitic or uninfected. While poor results were obtained using ResNet50 I achieved an accuracy of approx. 98% with validation accuracy about approx. 85% using VGG19 and achieved an accuracy of 99.44% with validation accuracy about 89.55% using CNN model made from scratch
Show More Show LessDescription
The comparison of India with other Countries during the initial phase of virus spread is visualized. Its comparison with initial hard-hit countries is also performed. Finally, prediction for time series analysis using prophet for the next seven days showing confirmed, recovered, and deaths that will happen was achieved.
Show More Show LessDescription
Flipkart web crawling and scrapping for product review sentiment analysis and deployment on Azure and Heroku. (GitHub) (Flipkart Scrapper - Azure) I combined both crawling and scraping for collecting reviews from Flipkart website. The end product Flipkart review scrapper is deployed on Azure and Heroku platforms click on web app above to see the project. Enter product name like Samsung or if you have specific requirements for product enter it like iphone7 (no spaces) on the search bar to get the reviews from the web site
Show More Show LessDescription
Downloading images one by one to collect a particular dataset is a hectic task and requires lot of manmade efforts for it. Here, I combined both crawling and scraping for collecting images from Google Images. These images are stored in jpg file format for further data analysis and processing
Show More Show LessDescription
To build a classification methodology to predict the type of forest cover based on the given training data. It is an AI-based web application, deployed in Heroku, GCP and Azure. This Web-based application has the capability to train itself with a single button click and predict results with good accuracy. CI CD pipeline has been followed while creating the project.
Show More Show LessDescription
To build a classification methodology to predict the type of thyroid based on the given training data. Data Analysis using python libraries is done for this project and machine learning models are compared and then on the best model hyper parameter tuning is done for prediction the type of thyroid. At the end the product as deployed on Azure cloud services.
Show More Show LessDescription
To build a regression model to predict the concrete compressive strength based on the different features in the training data. It is an AI-based web application, deployed in local system. This Web-based application has the capability to train itself with a single button click and predict results with good accuracy.
Show More Show LessDescription
To build a classification methodology to predict whether a costumer is going to be a defaulter for a payment or not. It is an AI-based web application, deployed in local system. This Web-based application has the capability to train itself with a single button click and predict results with good accuracy.
Show More Show LessDescription
To build a classification methodology to predict whether a website is a phishing website on the basis of given set of predictors. It is an AI-based web application, deployed in local system. This Web-based application has the capability to train itself with a single button click and predict results with good accuracy.
Show More Show LessDescription
To build a classification methodology to predict whether a website is a phishing website on the basis of given set of predictors. It is an AI-based web application, deployed in local system. This Web-based application has the capability to train itself with a single button click and predict results with good accuracy.
Show More Show LessDescription
The comparison of India with other Countries during the initial phase of virus spread is visualized. Its comparison with initial hard-hit countries is also performed. Finally, prediction for time series analysis using prophet for the next seven days showing confirmed, recovered, and deaths that will happen was achieved.
Show More Show LessDescription
I combined both crawling and scraping for collecting reviews from Flipkart website. The end product Flipkart review scrapper is deployed on Azure and Heroku platforms click on web app above to see the project. Enter product name like Samsung or if you have specific requirements for product enter it like iphone7 (no spaces) on the search bar to get the reviews from the web site.
Show More Show LessDescription
Downloading images one by one to collect a particular dataset is a hectic task and requires lot of manmade efforts for it. Here, I combined both crawling and scraping for collecting images from Google Images. These images are stored in jpg file format for further data analysis and processing.
Show More Show Less