Atsuo Kato's research while affiliated with Japan Research Institute and other places

Publications (11)

Article
Full-text available
We use a sparse variational dropout Bayesian neural network (SVDBNNs) to propose an investment strategy that gives consideration to predictive uncertainty. The proposed method is validated through simulation on historical orderbook data from the Tokyo Stock Exchange. Our results were found to outperform other standard non-Bayesian approaches on ris...
Article
Full-text available
While exchanges and regulators are able to observe and analyze the individual behavior of financial market participants through access to labeled data, this information is not accessible by other market participants nor by the general public. A key question, then, is whether it is possible to model individual market participants’ behaviors through...
Conference Paper
Financial markets are known to have difficulties in predicting, such as huge elements involved, unsteady internal structure, and existence of the market impact. Even when machine learning and deep learning methods are applied, predictions must include uncertainty, and investment decision making using uncertain prediction may cause large losses and...
Article
Full-text available
Prediction of financial market data with deep learning models has achieved some level of recent success. However, historical financial data suffer from an unknowable state space, limited observations, and the inability to model the impact of your own actions on the market can often be prohibitive when trying to find investment strategies using deep...
Conference Paper
Full-text available
We propose a scheme for selecting stocks related to a theme. This scheme was designed to support fund managers who are building themed mutual funds. Our scheme is a type of natural language processing method and based on words extracted according to their similarity to a theme using word2vec and our unique similarity based on co-occurrence in compa...
Chapter
Forecasting financial market trends is challenging. Predicting financial market trends always involves uncertainty because the economy is a complex system with a wide variety of interactions. Thus, to consider uncertainty, trends must be estimated stochastically. Conventional machine learning and deep learning methods cannot learn prediction uncert...
Conference Paper
Full-text available
Accurate prediction of financial markets is considered one of the most difficult problems due to the nature of its complexity, influenceability, and nonstationarity. Recent financial forecasting applications using neural networks typically have not taken the predictive uncertainty into consideration. Without proper consideration of predictive uncer...
Conference Paper
In recent years, predictions by machine learning and deep learning methods are utilized in various scenes of society. A model trained with deep learning methods can predict the target with high accuracy, but can not consider the predictive confidence sufficiently, and may predict high confident for extrapolated data which is hard to predict. In thi...
Conference Paper
Full-text available
本研究において,文書内における単語の共起を利用した上位下位概念の推定の手法を提案した.本手法に基づき,Wikipediaの記事を文書として使い,Wordnetに含まれる上位下位概念をデータとして実験を行った. その結果,精度は低いものの,一定の有効性を確認することができた.本手法は非常に少ない訓練データで必要なパラメータのチューニングが可能であることもわかった.
Article
Full-text available
We propose an extended scheme for selecting related stocks for themed mutual funds. This scheme was designed to support fund managers who are building themed mutual funds. In our preliminary experiments, building a themed mutual fund was found to be quite difficult. Our scheme is a type of natural language processing method and based on words extra...
Conference Paper
We propose a method to select and rank stocks related to a given theme. The proposed method has two flows; obtaining related words, and selecting related stocks based on obtained related words. First, on the basis of the given theme word, the proposed method selects words with high similarity using an ensemble of word2vec models. Then, we modify th...

Citations

... That way, we do not have to find complicated rules for agent behaviour, rather a utility function that agents want to optimize, using the actions available to them. This combination of methods was used successfully for various different systems [28][29][30][31][32][33][34][35]. A generic framework that makes use of this synergy was developed in [36] and expanded with an iterative learning approach in [37]. ...
... In [19], a Word2Vec model was used to learn embeddings of companies from news articles dataset. Ref. [20], Word2Vec model was trained using news and Wikipedia articles and companies' official disclosure files to identify embeddings for Japanese companies to select stocks for themed funds. To the best of our knowledge, the present work is the first work analyzing mutual funds filing data using NLP techniques. ...
... For example, [10], [18], [19] tackle stock returns forecasting using event embeddings obtained from financial news. [20] train company embeddings by applying BERT to the textual data from annual reports and [21] use word embeddings to select similar stocks. Despite applying the idea of embeddings within the financial domain, the aforementioned literature still relies on the aggregation of pretrained word embeddings from language models, rather than on using a novel technique to learn embeddings from non-textual financial data such as historical returns. ...