GitHub - DarthZhang/MachineLearning: Machine Learning project

Branches Tags

Name		Name	Last commit message	Last commit date
Latest commit History 58 Commits
test2		test2
train2		train2
.gitignore		.gitignore
ReadMe.txt		ReadMe.txt
X2.npz		X2.npz
cm_clas_report.py		cm_clas_report.py
config.py		config.py
data1.tar.xz		data1.tar.xz
data2.tar.xz		data2.tar.xz
data3.tar.xz		data3.tar.xz
data_cleaner.py		data_cleaner.py
ensemble.py		ensemble.py
exploratory.py		exploratory.py
gridsearch.py		gridsearch.py
gridsearch_full_label.py		gridsearch_full_label.py
heatmap.py		heatmap.py
liwc_test.csv		liwc_test.csv
liwc_train.csv		liwc_train.csv
load_data.py		load_data.py
load_data_from_zero.py		load_data_from_zero.py
lstm.py		lstm.py
mbti_1.csv		mbti_1.csv
oversample_full_label.py		oversample_full_label.py
predicting-personality-type.pdf		predicting-personality-type.pdf
smote.py		smote.py
split_types.py		split_types.py
svm.py		svm.py
train_test_split.npz		train_test_split.npz

Repository files navigation

Predicting Personality Type from Online Forum Text


The project was done in Python 3.6. Make sure to also install all the imported packages. 

The dataset used can be found here: https://www.kaggle.com/datasnaek/mbti-type

The main scripts are:
config.py - configure the path of the source dataset
lstm.py - uses a Long Short Term Memory neural network for classification
svm.py - uses a Support Vector Machine for classification
gridsearch.py, gridsearch_full_label.py - use gridsearch to find the best parameters
ensemble.py - uses an ensemble classifier
heatmap.py - generates the heatmap of the LIWC features