Data sets and trained ML model for ClickBait classifier
This MR consist of the following work:
-
Datasets for ClickBait Model. All data related to Click-Bait model is stored on Kaggle along with there notebook. link to dataset
-
Description of the dataset and Jupyter Notebook for data processing is included in this PR.
path to Data processing Notebook:
ML\Click-Bait\Utilis\CB_data_processing.ipynb
path to Data description:
ML\Click-Bait\Data\ClickBait_Data_Desc.md
-
Two models were trained:
-
simple tfidf vectorization as embedding and Multinomial Bayes classifier and simple Neural Network classifier: Accuracy 71%. link to notebook hosted on kaggle
-
using Google news Word2Vec as embedding and LSTM Neural Networks classifier: accuracy 82%. link to notebook hosted on kaggle
These Jupyter Notebooks are also included in this MR.
ML\Click-Bait\Models\Jupyter Notebook
-
-
all trained weights of these models, along with improvements, can be found on kaggle, as the output of these notebooks.
-
accuracy matrices are printed in notebooks.