gsmkbr
Experienced researcher in data science and its implementation in fluid mechanics problems, fault diagnosis, and various engineering/medical applications.
QIAUMecahnical Engineering Department, Qazvin Islamic Azad University, Qazvin, Iran
Pinned Repositories
Bearing_Fault_Diagnosis
This project attempts to identify nine different faults and the healthy condition based on vibrational dataset of bearings. For this purpose, the original time series signals are transformed into 2D wavelet maps. Due to few number of wavelet maps, a pretrained convolutional network (VGG16) is transfered to the deep neural network model. By freezing the initial layers and fine-tuning the last layers of VGG16 and fully-connected network, an accuracy higher than 94% is obtained for this 10-class project.
Compressor_Fault_Diagnosis
This project deals with detection and diagnosis of 7 faults of a reciprocating compressor using acoustic signals. After feature extraction, various machine learning algorithms are utilized to classify the faults.
Coursera_AI_Capstone_Project
This repostitory belongs to my final project in the course package "IBM AI Engineering"
ddpg_continuous_lunar_lander
Frailty_Classification_based_on_Kinect_Pose
This project deals with calssification of frailty level of humans based on Kinect time series obtained during various physical exercises.
gsmkbr
Config files for my GitHub profile.
ML_reconstruction_of_PIV_map
In this project, gappy regions of PIV vector maps are constructed by machine learning algorithms as well as popular statistical interpolation techniques.
RL_PPO_Cartpole
Reinforcement Learning based on Proximal Policy Optimization (PPO), on the Cartpole environment
gsmkbr's Repositories
gsmkbr/Compressor_Fault_Diagnosis
This project deals with detection and diagnosis of 7 faults of a reciprocating compressor using acoustic signals. After feature extraction, various machine learning algorithms are utilized to classify the faults.
gsmkbr/Bearing_Fault_Diagnosis
This project attempts to identify nine different faults and the healthy condition based on vibrational dataset of bearings. For this purpose, the original time series signals are transformed into 2D wavelet maps. Due to few number of wavelet maps, a pretrained convolutional network (VGG16) is transfered to the deep neural network model. By freezing the initial layers and fine-tuning the last layers of VGG16 and fully-connected network, an accuracy higher than 94% is obtained for this 10-class project.
gsmkbr/ML_reconstruction_of_PIV_map
In this project, gappy regions of PIV vector maps are constructed by machine learning algorithms as well as popular statistical interpolation techniques.
gsmkbr/Coursera_AI_Capstone_Project
This repostitory belongs to my final project in the course package "IBM AI Engineering"
gsmkbr/ddpg_continuous_lunar_lander
gsmkbr/Frailty_Classification_based_on_Kinect_Pose
This project deals with calssification of frailty level of humans based on Kinect time series obtained during various physical exercises.
gsmkbr/gsmkbr
Config files for my GitHub profile.
gsmkbr/RL_PPO_Cartpole
Reinforcement Learning based on Proximal Policy Optimization (PPO), on the Cartpole environment