ABSTRACT
This work aims to investigate the best methods for classifying textual information. As a dataset, we use advertisements with title and descriptions from the marketplace Jiji.ng who was partner for this work. The architecture is implemented on the basis of Deep Neural Networks such as CNNs and RNNs. All experiments were conducted using the Python programming language and framework for quick prototyping of Neural Networks - Keras.
The object of study is advertisements at e-commerce platform.
The subject of study is classification model for advertisements.
The aim of the study is to:
- find the best methods to represent words in vector space;
- study algorithms and methods for text classification;
- build a software for advertisements classification.
Key words: Natural Language Processing, Embeddings, CNN, RNN.