K.A

K.A

Finance, Data Science

Dark Mode

About me

Finance and Data Science.

Skills

  • Programming
    • Python
      • TensorFlow
      • Pytorch
      • OpenCV
    • R
    • SQL
  • Statistics
    • Machine Learning
    • Natural Language Processing
    • Time Series Analysis
    • Uplift Modeling
    • Bayesian Statistics
  • Finance
    • Financial Engineering
    • Duration Analysis

mcmc

Conference Presentations

  • Poster presentation on the 2019 International Conference on Data Science (Dec 15, 2019, Fudan University, China)
  • Keynote on Japanese Joint Statistical Meeting 2019 (September 8, 2019, Shiga University, Japan)
  • Poster presentation on the 13th Japan Statistical Society Spring Meeting (March 10, 2019, Nihon University, Japan)
  • Poster presentation on the 33rd Annual Conference of the Japanese Society for Artificial Intelligence (June 4 - 7, 2019, Niigata Convention Center, Japan).

Papers

①（Paper）

Asaba, K. (2019). Scam Cryptocurrency Detections using Machine Learning Techniques. In Proceedings of Japanese Society for Artificial Intelligence 33. 2019

  • Build models that receive cryptocurrencies’ whitepapers as input and predict whether the currencies are scam coin with Random Forest, Neural Net Classifier, and Decision Tree.
  • Achieved 90% accuracy with Random Forest. crypto

②（Paper・Presentation）

Stock Transaction Interval Analysis using Hawkes Process with Time-Varying Parameters

  • Analyzed and compared Japanese stocks’ transaction intervals with Hawkess process and ACD Model.
  • Confirmed significant difference between companies with different market cap group. hawkes

③（Paper）

Unsupervised Neural Network approaches for Morphological Analysis

  • Created models which can separate words into morphemes (e.g. worked → work + ed) in an unsupervised way using RNN, LSTM, Neural Network.
  • Achieved 91% accuracy with newly proposed attention model. hawkes