Tripadvisor Sentiment analysis on Six Senses Laamu

This repo is about detecting the emotion of people on visiting this amazing place and flag sentiment using scrapped data from Tripadvisor. A pie chart presented 99.3% of positive review and only 0.5% negative review from this analysis.

IMDB Sentiment analysis

The document presents the accuracy scores, confusion matrices, and classification reports for both Naive Bayes and Random Forest algorithms that is trained on IMDB dataset. It provides a function to predict sentiment based on input text. Any negative or positive review will be flagged correctly.

Rain Forecasting with Time series and ANN

This analysis has revealed some crucial factors of changing weather in Seattle. Based on the time series analysis, the average temperature has significantly increased over the years (1948-2017) in Seattle and the precipitation is still the same or even in a downward trend.

Churn Rate Prediction ANN Model

Churn prediction is the process of identifying, based on a customer's behaviour with product, which consumers are most likely to leave bank. ANN model has been employed for predicting which customer is most likely to leave the bank than those who most likely won't leave.

French to English machine translation - seq2seq LSTM architecture

I have made a Machine Translation model with Recurrence-based Seq2Seq architecture to translate French texts to English texts. The dataset I've used comes from Tatoeba, a collection of sentence translations in a variety of languages sourced from volunteers: https://tatoeba.org/en

Question Answering app

Hey there! I crafted a cool Q&A app using Streamlit and TensorFlow. I’m happy to share this Question Answering system app with Huggingface transformer that I worked on. Check it out here: https://lnkd.in/g29-Mxkn Also app link: https://lnkd.in/gcsXH9Xq

Modular Computer vision model

I have created a modular vision model, which can get commands and train as well as test the model with one line of command line code. The possibility is endless as you can also add other models on your own, use other arguments and the changes in code is endless.

Satellite map generation with pix2pix model

This project trains a GAN using Keras to translate satellite images into corresponding maps. The Pix2Pix model learns to create realistic maps by battling a discriminator that identifies real vs. generated maps. This has applications in urban planning, disaster response, and environmental monitoring