项目作者： mekhod

项目描述：

  The text data of 25000 labeled (positive and negative) IMDB reviews are processed and used to train a deep neural network model using Keras.

高级语言： HTML

项目主页：

项目地址: git://github.com/mekhod/imdb_rating.git

创建时间： 2017-10-31T03:56:55Z
项目社区：https://github.com/mekhod/imdb_rating
开源协议：
下载

IMDB_rating practice project

Project Overview:

In this practice project a labeled IMDB review dataset (positive or negative) is used to train a deep nueral network binary classifier.

The dataset is loaded and preprocessed using the Tokenizer method available in keras.preprocessing.text.
A deep neural network architecture (consisting of two fully connected dense layers and a sigmoid layer) is constructed.
The model is trained and the best model weights are stored.
The model with the best model weights are constructed and is evaluated against the test dataset.