Goal

The objective of the project was to improve my skills to use natural language processing techniques and to know how to apply that knowledge

Data preprocessing and cleaning

Removing noise and irrelevant data. Transforming text into a structured and processable format.

Data representation and preparation:

Vectorization: Converting text into numerical vectors that models can interpret. One-hot encoding: Creating binary representations of categorical data. Padding: Ensuring uniform sequence lengths for batch processing.

Model building with TensorFlow:

Designing and training models using Bidirectional LSTMs to capture context both forward and backward. Implementing neural network architectures adapted for sentiment analysis.

Machine learning and hyperparameter tuning:

Selecting optimal parameters to enhance model performance. Evaluating and fine-tuning the architecture to minimize error and maximize accuracy.

Result

I obtained a classifier that divides the data into negative positive and neutral and with this I understood the uses that can be given to it

Uses

• Identify positive negative or neutral sentiments in large datasets.
• Monitor the reputation of a brand, product, or service in real-time.
• Help make strategic decisions based on the behavior and emotions of the audience.
• Detect problems or points of improvement by analysing criticisms or complaints.
• Improve the personalization of services and products according to the customer's preferences and emotions.