Chuan-Ju Wang

Risk ranking from financial reports

This paper attempts to use soft information in finance to rank the risk levels of a set of companies. Specifically, we deal with a ranking problem with a collection of financial reports, in which each report is associated with a company. By using text information in the reports, which is so-called the soft information, we apply learning-to-rank techniques to rank a set of companies to keep them in line with their relative risk levels. In our experiments, a collection of financial reports, which are annually published by publicly-traded companies, is employed to evaluate our ranking approach; moreover, a regression-based approach is also carried out for comparison. The experimental results show that our ranking approach not only significantly outperforms the regression-based one, but identifies some interesting relations between financial terms.

On the construction and complexity of the bivariate lattice with stochastic interest rate models

Complex financial instruments with multiple state variables often have no analytical formulas and therefore must be priced by numerical methods, like lattice ones. For pricing convertible bonds and many other interest rate-sensitive products, research has focused on bivariate lattices for models with two state variables: stock price and interest rate. This paper shows that, unfortunately, when the interest rate component allows rates to grow in magnitude without bounds, those lattices generate invalid transition probabilities. As the overwhelming majority of stochastic interest rate models share this property, a solution to the problem becomes important. This paper presents the first bivariate lattice that guarantees valid probabilities. The proposed bivariate lattice grows (super)polynomially in size if the interest rate model allows rates to grow (super)polynomially. Furthermore, we show that any valid constant-degree bivariate lattice must grow superpolynomially in size with log-normal interest rate models, which form a very popular class of interest rate models. Therefore, our bivariate lattice can be said to be optimal.

ICE: Item Concept Embedding via Textual Information

This paper proposes an item concept embedding (ICE) framework to model item concepts via textual information. Specifically, in the proposed framework there are two stages: graph construction and embedding learning. In the first stage, we propose a generalized network construction method to build a network involving heterogeneous nodes and a mixture of both homogeneous and heterogeneous relations. The second stage leverages the concept of neighborhood proximity to learn the embeddings of both items and words. With the proposed carefully designed ICE networks, the resulting embedding facilitates both homogeneous and heterogeneous retrieval, including item-to-item and word-to-item retrieval. Moreover, as a distributed embedding approach, the proposed ICE approach not only generates related retrieval results but also delivers more diverse results than traditional keyword-matching-based approaches. As our experiments on two real-world datasets show, ICE encodes useful textual information and thus outperforms traditional methods in various item classification and retrieval tasks.

Financial Keyword Expansion via Continuous Word Vector Representations

This paper proposes to apply the continuous vector representations of words for discovering keywords from a financial sentiment lexicon. In order to capture more keywords, we also incorporate syntactic information into the Continuous Bag-ofWords (CBOW) model. Experimental results on a task of financial risk prediction using the discovered keywords demonstrate that the proposed approach is good at predicting financial risk.

A Closed-Form Formula for an Option with Discrete and Continuous Barriers

This article presents a methodology to derive analytical formulas for a class of complicated financial derivatives with a continuously monitored barrier and a few discretely monitored ones. Numerical results based on concrete numbers for the parameters are presented and analyzed.

On the risk prediction and analysis of soft information in finance reports

We attempt in this paper to utilize soft information in financial reports to analyze financial risk among companies. Specifically, on the basis of the text information in financial reports, which is the so-called soft information, we apply analytical techniques to study relations between texts and financial risk. Furthermore, we conduct a study on financial sentiment analysis by using a finance-specific sentiment lexicon to examine the relations between financial sentiment words and financial risk. A large collection of financial reports published annually by publicly-traded companies is employed to conduct our experiments; moreover, two analytical techniques – regression and ranking methods – are applied to conduct these analyses. The experimental results show that, based on a bag-of-words model, using only financial sentiment words results in performance comparable to using the whole texts; this confirms the importance of financial sentiment words with respect to risk prediction. In addition to this performance comparison, via the learned models, we draw attention to some strong and interesting correlations between texts and financial risk. These valuable findings yield greater insight and understanding into the usefulness of soft information in financial reports and can be applied to a broad range of financial and accounting applications.

Discovering Finance Keywords via Continuous-Space Language Models

The growing amount of public financial data makes it increasingly important to learn how to discover valuable information for financial decision making. This article proposes an approach to discovering financial keywords from a large number of financial reports. In particular, we apply the continuous bag-of-words (CBOW) model, a well-known continuous-space language model, to the textual information in 10-K financial reports to discover new finance keywords. In order to capture word meanings to better locate financial terms, we also present a novel technique to incorporate syntactic information into the CBOW model. Experimental results on four prediction tasks using the discovered keywords demonstrate that our approach is effective for discovering predictability keywords for post-event volatility, stock volatility, abnormal trading volume, and excess return predictions. We also analyze the discovered keywords that attest to the ability of the proposed method to capture both syntactic and contextual information between words. This shows the success of this method when applied to the field of finance.

Deep belief networks for predicting corporate defaults

This paper provides a new perspective on the default prediction problem using deep learning algorithms. Via the advantages of deep learning, the representable factors of input data will no longer need to be explicitly extracted, but can be implicitly learned by the deep learning algorithms. We consider the stock returns of both default and solvent companies as input signals and adopt one of the deep learning architecture, Deep Belief Networks (DBN), to train the prediction models. The preliminary results show that the proposed approach outperforms traditional machine learning algorithms.

Evaluating corporate bonds with complicated liability structures and bond provisions

This paper presents a general and numerically accurate lattice methodology to price risky corporate bonds. It can handle complex default boundaries, discrete payments, various asset sales assumptions, and early redemption provisions for which closed-form solutions are unavailable. Furthermore, it can price a portfolio of bonds that accounts for their complex interaction, whereas traditional approaches can only price each bond individually or a small portfolio of highly simplistic bonds. Because of the generality and accuracy of our method, it is used to investigate how credit spreads are influenced by the bond provisions and the change in a firm’s liability structure due to bond repayments.

Post-Modern Portfolio Theory for Information Retrieval

Abstract Information Retrieval (IR) aims to discover relevant information according to a users information need. The classic Probability Ranking Principle (PRP) forms the theoretical basis for probabilistic IR models. This ranking principle, however, neglects the uncertainty introduced through the estimations from retrieval models. Inspired by the Post-Modern Portfolio Theory (PMPT), this paper proposes a mean-semivariance framework to handle the uncertainty. The proposed framework not only deals with the uncertainty but has the ability to distinguish bad surprises (downside uncertainty) and good surprises (upside uncertainty) when optimizing a ranking list. The experimental results shows that the proposed method improves the IR performance over the PRP baseline in terms of most of IR evaluation metrics; moreover, the results suggest that the mean-semivariance framework can further boost the top-position ranking quality.