All Projects → baiyyang → medical-diagnosis-cnn-rnn-rcnn

baiyyang / medical-diagnosis-cnn-rnn-rcnn

Licence: Apache-2.0 License
分别使用rnn/cnn/rcnn来实现根据患者描述,进行疾病诊断

Programming Languages

python
139335 projects - #7 most used programming language

Projects that are alternatives of or similar to medical-diagnosis-cnn-rnn-rcnn

Multi Class Text Classification Cnn Rnn
Classify Kaggle San Francisco Crime Description into 39 classes. Build the model with CNN, RNN (GRU and LSTM) and Word Embeddings on Tensorflow.
Stars: ✭ 570 (+1361.54%)
Mutual labels:  text-classification, lstm, rnn
Textclassificationbenchmark
A Benchmark of Text Classification in PyTorch
Stars: ✭ 534 (+1269.23%)
Mutual labels:  text-classification, lstm, rcnn
automatic-personality-prediction
[AAAI 2020] Modeling Personality with Attentive Networks and Contextual Embeddings
Stars: ✭ 43 (+10.26%)
Mutual labels:  text-classification, lstm, rnn
Pytorch Rnn Text Classification
Word Embedding + LSTM + FC
Stars: ✭ 112 (+187.18%)
Mutual labels:  text-classification, lstm, rnn
sequence-rnn-py
Sequence analyzing using Recurrent Neural Networks (RNN) based on Keras
Stars: ✭ 28 (-28.21%)
Mutual labels:  lstm, rnn
SpeakerDiarization RNN CNN LSTM
Speaker Diarization is the problem of separating speakers in an audio. There could be any number of speakers and final result should state when speaker starts and ends. In this project, we analyze given audio file with 2 channels and 2 speakers (on separate channels).
Stars: ✭ 56 (+43.59%)
Mutual labels:  lstm, rnn
deep-improvisation
Easy-to-use Deep LSTM Neural Network to generate song sounds like containing improvisation.
Stars: ✭ 53 (+35.9%)
Mutual labels:  lstm, rnn
yunyi
2018“云移杯- 景区口碑评价分值预测
Stars: ✭ 29 (-25.64%)
Mutual labels:  text-classification, rnn
Paper-Implementation-DSTP-RNN-For-Stock-Prediction-Based-On-DA-RNN
基於DA-RNN之DSTP-RNN論文試做(Ver1.0)
Stars: ✭ 62 (+58.97%)
Mutual labels:  lstm, rnn
theano-recurrence
Recurrent Neural Networks (RNN, GRU, LSTM) and their Bidirectional versions (BiRNN, BiGRU, BiLSTM) for word & character level language modelling in Theano
Stars: ✭ 40 (+2.56%)
Mutual labels:  lstm, rnn
Speech-Recognition
End-to-end Automatic Speech Recognition for Madarian and English in Tensorflow
Stars: ✭ 21 (-46.15%)
Mutual labels:  lstm, rnn
air writing
Online Hand Writing Recognition using BLSTM
Stars: ✭ 26 (-33.33%)
Mutual labels:  lstm, rnn
novel writer
Train LSTM to writer novel (HongLouMeng here) in Pytorch.
Stars: ✭ 14 (-64.1%)
Mutual labels:  lstm, rnn
lstm-electric-load-forecast
Electric load forecast using Long-Short-Term-Memory (LSTM) recurrent neural network
Stars: ✭ 56 (+43.59%)
Mutual labels:  lstm, rnn
Customer-Feedback-Analysis
Multi Class Text (Feedback) Classification using CNN, GRU Network and pre trained Word2Vec embedding, word embeddings on TensorFlow.
Stars: ✭ 18 (-53.85%)
Mutual labels:  text-classification, rnn
OCR
Optical character recognition Using Deep Learning
Stars: ✭ 25 (-35.9%)
Mutual labels:  lstm, rcnn
question-pair
A siamese LSTM to detect sentence/question pairs.
Stars: ✭ 25 (-35.9%)
Mutual labels:  lstm, rnn
ConvLSTM-PyTorch
ConvLSTM/ConvGRU (Encoder-Decoder) with PyTorch on Moving-MNIST
Stars: ✭ 202 (+417.95%)
Mutual labels:  lstm, rnn
object-tracking
Multiple Object Tracking System in Keras + (Detection Network - YOLO)
Stars: ✭ 89 (+128.21%)
Mutual labels:  lstm, rcnn
ArrayLSTM
GPU/CPU (CUDA) Implementation of "Recurrent Memory Array Structures", Simple RNN, LSTM, Array LSTM..
Stars: ✭ 21 (-46.15%)
Mutual labels:  lstm, rnn

medical-diagnosis-cnn-rnn-rcnn

分别使用rnn/cnn/rcnn模型来实现根据患者描述,进行疾病诊断

Describe

分别使用rnn,cnn,rcnn模型来实现医疗疾病诊断,即文本分类工作

  1. rnn模型如下,经过LSTM取最后一刻的输出,之后经过softmax函数分类,模型的结构如下图所示。
  2. cnn模型参考Yoon Kim论文《Convolutional Neural Networks for Sentence Classification》实现,模型结构图如下所示。

通过实验研究发现,由于该网络中,只取最大的权重作为最后的分类,容易存在过拟合。因此,我们改进了原始的TextCNN,提出了:

  • 基于平均特征层的卷积神经网络(Mean Features Convolutional Nerual Network for Sentence Classification,MF-TextCNN)
  • 基于全特征相连层下的卷积神经网络文本分类模型(All Features Concat Convolutional Nerual Network for Sentence Classification,AFC-TextCNN)
  1. rcnn为Bi-LSTM后接高度为1,宽度为2*hidden_dim的卷积核,之后取max-pooling,最后经过softmax函数。模型的结构如图所示。

Requirements

  • python 3
  • tensorflow >= 1.5
  • numpy
  • zhon
  • jieba

Result

models presicion note
rnn 0.78 收敛较慢
textcnn 0.82 收敛较快,容易过拟合
MF-textcnn 0.80 收敛较慢,欠拟合
AFC-textcnn 0.88 收敛速度一般,准确率较高
rcnn 0.84 收敛较快

Others

欢迎各位大佬指正!

Note that the project description data, including the texts, logos, images, and/or trademarks, for each open source project belongs to its rightful owner. If you wish to add or remove any projects, please contact us at [email protected].