Embedded Files
Ethani Caphace
Highlights
Highlights
Amharic Speech Recognition
Amharic Speech Recognition
Deep Learning has changed the game when it comes to voice recognition by introducing end-to-end models. These models take in an audio signal and directly output transcriptions.
AMHARIC SPEECH RECOGNITION
AMHARIC SPEECH RECOGNITION
Data Pre-processing:
Data Pre-processing:
Perform audio data preprocessing like resampling, normalization, and removing noise.
Perform data augmentation by adding noise, changing the speed and pitch of the audio.
Implemented audio data feature extraction method using log Mel spectrogram.
Machine Learning:
Machine Learning:
Build a model with have two main neural network modules. Three layers of Residual Convolutional Neural Networks to learn the relevant audio features. Then a set of Bidirectional Recurrent Neural Networks to leverage the learned audio features.
Pharmaceutical Sales prediction across multiple stores.
Pharmaceutical Sales prediction across multiple stores.
A regression problem, whose main aim is to come up with an end-to-end product that delivers Sales prediction across multiple stores of some Pharmaceutical companies. The performance of 3 regression models is explored: Linear Regression, XGBoost, and Random Forest. Random Forest regressor emerges as the best performer with a Mean Square Error of 0.056. Streamlit is used for model deployment and visualization.
A regression problem, whose main aim is to come up with an end-to-end product that delivers Sales prediction across multiple stores of some Pharmaceutical companies. The performance of 3 regression models is explored: Linear Regression, XGBoost, and Random Forest. Random Forest regressor emerges as the best performer with a Mean Square Error of 0.056. Streamlit is used for model deployment and visualization.
APPROACH
EDA:
Data exploration using Pandas, Matplotlib, Numpy modular code.
Creation of new features.
Machine Learning:
Machine Learning:
Applied Random Forest model that takes multiple variables as input and predicts sales
Applied an LSTM recurrent neural network that takes seven weeks of historical sales data and makes predictions for future sales.
Calculated Feature Importance to see which variables are mainly responsible for affecting Sale and number.
Telecommunication Users Data Analysis.
Telecommunication Users Data Analysis.
Using Telecommunication's data to perform Data Exploratory, Customers Overview, User Engagement, Experience and Satisfaction Analysis.
Using Telecommunication's data to perform Data Exploratory, Customers Overview, User Engagement, Experience and Satisfaction Analysis.
APPROACH
EDA:
Telecommunication Data exploration using Pandas, Matplotlib, Scikit-learn, Numpy, seaborn, and other common Data Exploratory packages.
The following are the tasks that have been performed,
1. Data Cleaning, Transformations and Feature Extractions,
2. Customers Overview,
3. Uni-variate and Multi-variate feature analysis,
4. Correlation Analysis,
5. User Engagement Analysis,
6. User Experience and Satisfaction Analysis.
Page updated
Google Sites
Report abuse