Image Data

Deep learning is an important technology for image recognition. Here we mention two interesting applications.

Deepmind Computer Is World’s Best Go Player #

Go is a board game invented in China more than 2500 years ago and it is played to the present day. The following figure shows the first 50 steps in the first match between AlphaGo and the world’s number one human player Ke Jie in May 2017.

Alpha Go VS Ke Jie

Silver et al. (2016) recognize the state of the game by images (as shown above) and predict the best action using deep learning methods:

Recently, deep convolutional neural networks have achieved unprecedented performance in visual domains: for example, image classification, face recognition, and playing Atari games. They use many layers of neurons, each arranged in overlapping tiles, to construct increasingly abstract, localized representations of an image. We employ a similar architecture for the game of Go.

See also Silver et al. (2017) for the improved algorithm that won the matches with Ke Jie.

In finance, OHLC chart shows the daily open, high, low, and closing prices for stocks as illusrated below.

OHLC Chart

Here is an example of the OHLC charts for Tesla extracted from Yahoo Finance.

OHLC Tesla

In the same figure you could also find:

  • A purple line showing the moving average prices
  • The volume bars at the bottom showing the volumn of trading.

Jiang, Kelly and Xiu (2022+, forthcoming at Journal of Finance) removed the colors and constructed their OHLC images to predict stock returns. Below shows an image instance from their dataset:

OHLC Image

which includes four variants with/without volume bars (VB) and with/without the 20-days moving average line (MA).

Previous Section: Why ReLU Function
Next Section: Spatial Information