Skip to content

aadishjain2911/Moodify

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

8 Commits
 
 

Repository files navigation

Moodify

Description

Have you ever experienced that the instagram posts visible to you are completely unrelated to what you wanted to see ? Well, not anymore. Moodify will classify your insta feed according to your current mood, which will make your feed more engrossing and relatable. We will use Instagram API for gathering the feed and apply our model to select relevant posts. We plan to build a ML model which will classify the posts based on their picture, caption, and top few comments to predict which mood is more likely to relate with this post. Initially we will learn a CNN based model for the image part and a RNN based model for the textual part. Then we will combine their outputs to predict a common metric and classify the outputs in "happy", "funny", "motivational" etc. We will only be working on those posts with a single image (in case of multiple images we will select the first one). Later, if time permits, we may also build a mobile application for the same.

Checkpoints

  • Learning basics of CNN and RNN
  • Understanding MetaForDevelopers API offered by instagram
  • Building the CNN model for image classification
  • Implementing RNN model for textual data
  • Combining both the models to get the final output, and hyper parameter tuning for achieving better results

Weekly Plan

Week 1

Reading material for this week -

Week 2

Reading material for this week -

Week 3

Reading material for this week -

Week 4

Read and understand how will we use the Meta For Developers API for our purpose i.e. extracting the insta feed of a person.

Week 5

Look up for dataset online for both image and textual data with annotated sentiments. Couple of examples -

Week 6

Build a basic pipeline of the model including input preprocessing, a very basic naive model, correct output format, and metric calculation. Don't pay attention to model details or metric values now.

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors