Research: Implementation of Cross Domain Recommender System (Research) (Open)

Dataset: Amazon product dataset 2018

Implementation platform: Google Colab/ Supercomputer

    Technology and Domain: Python(basic, numpy, pandas, tensorflow, keras), ML

Description: The users are active in some domain and inactive in some domain. The prediction will be difficult for the users when users are inactive. In this scenario, the RS can take help of other domain in which users are active and enhace the RS prediction. The other meta information can be used to further enhance the recommender system.

Outcome: By this research, we can check the usefulness of auxiliary domain for target domain prediction in RS. This work can be presented in conference/Journal.

Project Team Member/s: Two students required

Apply: forms.gle/N2XT6sAAz9Z6awSMA 

Research: Implementation of BERT Algorithm for enhancing product recommendation (Open)

Dataset: Amazon product dataset 2018

Implementation platform: Google Colab/ Supercomputer

    Technology and Domain: Python(basic, numpy, pandas, tensorflow, keras), ML

Description: The BERT algorithm has shown tremendous success for the NLP task. The same algorithm performance can be checked for the product recommender system. The reviews and descriptions are available for the item. This review is help full to enhance RS performance. The transfer learning was not possible before BERT algorithm. BERT has opened up this opportunity which can be compared with existing state of the art methods.

Outcome: By this research, we can check the transfer learning opportunity in RS using BERT. This work can be presented in conference.

Project Team Member/s: Two students required

Apply: forms.gle/N2XT6sAAz9Z6awSMA 

Project:8th semester project reporting and marking(Open)

Technology: PHP and My SQL or Any language to build website

Description: The requirements are gathered and wireframes are ready. The students need to built and host the website.

Project Team Members:

1) 

2) 

Apply: forms.gle/N2XT6sAAz9Z6awSMA 

Inspection Data Capture System  (Consultancy)

Project Modules: (1) Set Design Parameter (2)  Fetch Item Detail (3) Inspection Approval / Rejection (4) Recording Inspection detail (5) Report Download

Project Description: The QA process for the new items, uses lot of paperwork, just to inspect the values. This waste lots of paper and do not have the digital data. The past digital data helps to analyses the pattern and can take efficient decision on it. The current method records inspected value of an item in the hard copy. Inspection Data Capture System (IDCS) will help to address those problems. It will reduce humans’ efforts by easy testing parameter setup, use old parameter setup, linking with testing items, easy statical analysis and report generation.

Project Status:  Completed

Agency name:  Rotomag Motors & Control Pvt. Ltd.

Project Team Members:

1) Nimesh Italiya (17CE036)

2) Hardev Khandhar (18CE043)

3) Prof. Ronak Patel, CE Dept.

4) Dr. Ritesh Patel, CE Dept

Internship Company Approval App (Project)

Technology: Android and firebase

Project Team Members:

1) Katha Patel (17CE073)

2) Chandresh Mendapara (17CE057)

3) Pratik Dhoriyani (18CE024)

Enhancing recommender system with season and style features (Research).

Technology: Python, Keras

Project Team Members:

1) Raghav Modi (18CE056)

Paper Link: https://link.springer.com/chapter/10.1007/978-981-16-6309-3_44

Transfer learning in Image based product recommender system (Research)

Technology: Python, Keras

Project Team Members:

1) Shyamal Shah (D19CE152)

Paper Link: https://link.springer.com/chapter/10.1007/978-981-19-0098-3_8

Product Rating prediction from Review (Research)

In eCommerce the product reviews are available.  The review is taken and based on this given review what will be the rating of the product will be calculated by the RS. This rating is useful to find out the sentiment of users' for the product.


Image Based Product Recommender System (Research)

The product image which you like, can  be given to the RS  as an input and based on the image features it recommend you top-5 most similar products. The pre-trained CNN models are used to extract the visual features and NN algorithm is used to find out similar products.


Enhancing Movie Recommender System  (M.Tech Dissertation)

Recommender Systems are used by customers to buy items more efficiently. Business is also benefited simultaneously. There are various approaches of Recommender Systems, like: (1) collaborative Filtering (2) Content based Filtering and (3) Hybrid Filtering. We model Hybrid approach and make prediction task as classification problem where our aim is to predict whether the Movie will be liked or disliked by the user. We propose item based recommender which combines usage, tag and movie specific data such as genres, star cast and directors to improve the accuracy of the Recommender System. Tool: Matlab   Dataset: hetrec2011-movielens-2k


Geo spatial Feature based Geo-Metric Network Management System (Project)

(B.Tech 8th Semester project  at BISAG,  Gandhinagar)

It is based on GIS (Geo Graphic Information System). We have to display the image from shapefile of any road of the world which is in binary form and after that we locate two points on that image and find Shortest distance between  them by Dijkstra’s Algorithm. Language: C#  Tool: MS Visual Studio 2008


Web Usage Mining (Project)

Application of data mining techniques to the World Wide Web referred as Web Mining. We follow three basic steps of WUM for finding Association Rule which satisfy Minimum support and Confidence. We find Association Rule from log file which is generated at server side.  Three Steps are as follows: (1)Pre-processing (2)Pattern Discovery (3)Pattern Analysis Language: C#    Tool: MS Visual Studio 2008