marker - associations - binary - base Open Source Model - Free for Biomedical Text Marker Association Classification

Marker Associations Binary Base

Developed by jambo

A binary classification model fine-tuned on biomedical text using PubMedBERT, designed for marker association classification tasks

Text Classification

Transformers

Open Source License:MIT #Biomedical Text Classification #High Recall Rate #PubMedBERT Fine-tuning

Downloads 16

Release Time : 3/2/2022

Model Overview

This model is a fine-tuned version of microsoft/BiomedNLP-PubMedBERT-base-uncased-abstract-fulltext, specifically designed for handling binary classification tasks of marker associations in the biomedical field. It can identify association relationships between entities such as genes and chemical substances.

Model Features

Biomedical Domain Optimization

Optimized for biomedical text based on the PubMedBERT pre-trained model

High Recall Rate

Demonstrates high recall characteristics on the evaluation set, with a 100% recall rate for chemical substance categories

Balanced Performance

Achieves a good balance between precision and recall, with an F1 score around 0.87

Model Capabilities

Biomedical Text Classification

Gene Association Recognition

Chemical Substance Association Recognition

Binary Classification Task Processing

Use Cases

Biomedical Research

Gene-Disease Association Analysis

Identify association relationships between genes and diseases in literature

Precision 0.808, Recall 0.940

Drug-Target Interaction Identification

Extract interaction relationships between chemical substances and biological targets from literature

Precision 0.774, Recall 1.0

Literature Mining

Biomedical Entity Relationship Extraction

Extract association relationships between entities from biomedical literature such as PubMed

🚀 marker-associations-binary-base

This model is a fine - tuned version of [microsoft/BiomedNLP - PubMedBERT - base - uncased - abstract - fulltext](https://huggingface.co/microsoft/BiomedNLP - PubMedBERT - base - uncased - abstract - fulltext) on the marker - associations - binary - base dataset. It aims to solve text classification tasks on relevant medical data, providing high - precision classification results.

🚀 Quick Start

This model is ready to be used for text classification tasks. You can fine - tune it further according to your specific requirements.

✨ Features

Fine - tuned on the marker - associations - binary - base dataset, suitable for text classification in relevant medical scenarios.
Achieved good performance on evaluation metrics such as precision, recall, F1 - score, and accuracy.

📚 Documentation

Evaluation Results

This model achieves the following results on the evaluation set:

Gene Results

Precision = 0.808
Recall = 0.940
F1 = 0.869
Accuracy = 0.862
AUC = 0.944

Chemical Results

Precision = 0.774
Recall = 1.0
F1 = 0.873
Accuracy = 0.926
AUC = 0.964

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

learning_rate: 5e - 05
train_batch_size: 16
eval_batch_size: 16
seed: 1
optimizer: Adam with betas=(0.9,0.999) and epsilon = 1e - 08
lr_scheduler_type: linear
num_epochs: 15

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1	Accuracy	Auc
No log	1.0	88	0.3266	0.8191	0.8462	0.8324	0.8670	0.9313
No log	2.0	176	0.3335	0.7870	0.9341	0.8543	0.8755	0.9465
No log	3.0	264	0.4243	0.7982	0.9560	0.87	0.8884	0.9516
No log	4.0	352	0.5388	0.825	0.7253	0.7719	0.8326	0.9384
No log	5.0	440	0.7101	0.8537	0.7692	0.8092	0.8584	0.9416
0.1824	6.0	528	0.6175	0.8242	0.8242	0.8242	0.8627	0.9478

Framework versions

Transformers 4.11.3
Pytorch 1.9.0+cu111
Tokenizers 0.10.3

📄 License

This project is under the MIT license.

Featured Recommended AI Models

Empowering the Future, Your AI Solution Knowledge Base

English 简体中文繁體中文にほんご