-
Developed LSTM model for generating song lyrics.
​
-
Utilized songdata.csv for dataset, applying tokenization and sequence standardization
-
Implemented Bidirectional LSTM architecture for improved pattern learning
-
Achieved 70% accuracy on validation dataset (more accuracy can be reached with more no.of epochs and more layers in the model)
-
Generated coherent lyrics autonomously from simple starting words
-
Created a deep learning model for analysing the reviews for sentiment polarity (Positive or Negative).
​
-
Attained an accuracy of over 80%.
​
-
Used Streamlit for creating a webapp and deployed it for global use
​
-
It can be used to evaluate the sentiment of a movie review, product review and more.
​
​
-
Developed a predictive model to identify diabetes occurrences using Kaggle dataset.
-
Conducted Exploratory Data Analysis (EDA) to understand dataset characteristics.
-
Implemented K-Nearest Neighbors (KNN) algorithm for classification.
-
Achieved a robust 79% accuracy, showcasing model efficacy in predicting diabetes.
-
Model potential for aiding healthcare professionals in early intervention strategies.
-
Implemented random forest classifier to predict equipment failures with high accuracy.
-
Utilized generated data series with numpy to simulate failure rates for detailed pattern analysis.
-
Conducted meticulous data preprocessing and feature engineering to extract relevant features.
-
Achieved impressive accuracy score of 94% in predicting equipment failures.
-
Demonstrated proficiency in applying data-driven approaches to proactively manage equipment, enhancing operational efficiency in industrial settings.