Pannawit Samatthiyadikun

Pannawit Samatthiyadikun

Pannawit Samatthiyadikun

Why not create your own Wantedly profile?

Showcase yourself and your experiences to bring your professional story to life.

Join

The scientist with a research background in data analysis and mining, probabilistic and statistical modeling as well as similar prediction and convex optimization techniques for building recommender systems, and NLP application. Seeking growth in career path and to apply and further gains skills and knowledge within a commercially focused technical environment.

In the future

Ambition

In the future

My principal research interests lie in the field of data analysis and mining, probabilistic and statistical modeling as well as similar prediction and convex optimization techniques for building recommender systems. I also am interested in using new approaches, techniques, and technologies such as big data, deep learning, stream data, onl

Sept 2020

総合研究大学院大学

情報工学系

Sept 2020

PhD * Already finished all PhD course and waiting for the journal publishing process. Rather than wasting time waiting, I decided to work as a full-time employee and currently obtaining a working Visa.

Multicriteria collaborative filtering by Bayesian model-based user profiling

Aug 2012

Multicriteria collaborative filtering by Bayesian model-based user profiling

Aug 2012

Bayesian model for a multi-criteria recommender system with support vector regression

Aug 2013

Bayesian model for a multi-criteria recommender system with support vector regression

Aug 2013

Extended Bayesian Model for Multi-criteria Recommender System

Jan 2013

Extended Bayesian Model for Multi-criteria Recommender System

Jan 2013

Item Age Effect Against Global User Preference in Latent Model of Recommendation System

Aug 2015

Item Age Effect Against Global User Preference in Latent Model of Recommendation System

Aug 2015

Latent Models for Movie Recommendation System

As a part of five-year Ph.D. course, dissertation progress evaluation is performed in the second year and equivalent to Master’s thesis defense. I extended proposed the Bayesian probabilistic model, with an assumption the movies popularity is as decreased as they are getting older. The model is able to group up the same characteristics such as rating patterns by their age, and shows that not all movies go along with the stated assumption.

Latent Models for Movie Recommendation System

As a part of five-year Ph.D. course, dissertation progress evaluation is performed in the second year and equivalent to Master’s thesis defense. I extended proposed the Bayesian probabilistic model, with an assumption the movies popularity is as decreased as they are getting older. The model is able to group up the same characteristics such as rating patterns by their age, and shows that not all movies go along with the stated assumption.

Polylingual Topic Modeling for Scholarly Information Recommendation

This PhD dissertation is a study of Bayesian topic modeling on academic articles bibliography data with multiple features, and application examples of the model such as keyword recommendation. It is also extended to the special case that those articles are polylingual. We further investigate for improvement, especially using deep learning techniques, and integrated them into the model learning procedure with cutting-edge technique to boost up each topic topical quality and meaning.

Polylingual Topic Modeling for Scholarly Information Recommendation

This PhD dissertation is a study of Bayesian topic modeling on academic articles bibliography data with multiple features, and application examples of the model such as keyword recommendation. It is also extended to the special case that those articles are polylingual. We further investigate for improvement, especially using deep learning techniques, and integrated them into the model learning procedure with cutting-edge technique to boost up each topic topical quality and meaning.

Nov 2017
-
May 2019

Arblet Inc

2 years

Data Scientist

Nov 2017 - May 2019

· Expand self-knowledge to new areas such as anatomy, medical-related, signal processing, time series analysis, and other related research. · Design and execute the experiments for multiple projects such as defining an objective, necessary sufficient data and planning on the collection, and ensure an effect of confounding

Oct 2012
-
Aug 2017

リサーチアシスタント(Intern)

Oct 2012 - Aug 2017

· Implement the text models and other well-known models for accuracy comparison for lab internal usage for academic papers publishing. · Perform iterative experiment process; hyperparameter optimization, result analysis, and model accuracy improvement. · Provide meaningful visualization for further analysis and improvement.

Sept 2014

総合研究大学院大学

情報工学系

Sept 2014

MSc


Skills and qualities

Time Series

0

SVM

0

TensorFlow

0

Python

0

Machine Learning

0

Publications

Item Age Effect Against Global User Preference in Latent Model of Recommendation System

Aug 2015

Bayesian model for a multi-criteria recommender system with support vector regression

Aug 2013

Extended Bayesian Model for Multi-criteria Recommender System

Jan 2013

Multicriteria collaborative filtering by Bayesian model-based user profiling

Aug 2012

Accomplishments/Portfolio

Polylingual Topic Modeling for Scholarly Information Recommendation

Latent Models for Movie Recommendation System

Awards and Certifications

TOEIC 735 score

Aug 2013


Languages

English - Professional, Japanese - Conversational, Thai - Native