Nowadays, traditional machine learning methods for building predictive models of credit card customer churn are no longer sufficient for effective customer management. Additionally, interpreting these models has become essential. This study aims to balance the data using sampling techniques to forecast whether a customer will churn, combine machine learning methods to build a comprehensive customer churn prediction model, and select the model with the best performance. The optimal model is then interpreted using the Shapley Additive exPlanations (SHAP) values method to analyze the correlation between each independent variable and customer churn. Finally, the causal impacts of these variables on customer churn are explored using the R-learner causal inference method. The results show that the complete customer churn prediction model using Extreme Gradient Boosting (XGBoost) achieved significant performance, with accuracy, precision, recall, F1 score, and area under the curve (AUC) all reaching 97%. The SHAP values method and causal inference method demonstrate that several variables, such as the customer's total number of transactions, the total transaction amount, the total number of bank products, and the changes in both the amount and the number of transactions from the fourth quarter to the first quarter, have an impact on customer churn, providing a theoretical foundation for customer management.

Nowadays, the increasing number of medical diagnostic data and clinical data provide more complementary references for doctors to make diagnosis to patients. For example, with medical data, such as electrocardiography (ECG), machine learning algorithms can be used to identify and diagnose heart disease to reduce the workload of doctors. However, ECG data is always exposed to various kinds of noise and interference in reality, and medical diagnostics only based on one-dimensional ECG data is not trustable enough. By extracting new features from other types of medical data, we can implement enhanced recognition methods, called multimodal learning. Multimodal learning helps models to process data from a range of different sources, eliminate the requirement for training each single learning modality, and improve the robustness of models with the diversity of data. Growing number of articles in recent years have been devoted to investigating how to extract data from different sources and build accurate multimodal machine learning models, or deep learning models for medical diagnostics. This paper reviews and summarizes several recent papers that dealing with multimodal machine learning in disease detection, and identify topics for future research.

Volatility, a pivotal factor in the financial stock market, encapsulates the dynamic nature of asset prices and reflects both instability and risk. A volatility quantitative investment strategy is a methodology that utilizes information about volatility to guide investors in trading and profit-making. With the goal of enhancing the effectiveness and robustness of investment strategies, our methodology involved three prominent time series models with six machine learning models: K-nearest neighbors, AdaBoost, CatBoost, LightGBM, XGBoost, and random forest, which meticulously captured the intricate patterns within historical volatility data. These models synergistically combined to create eighteen novel fusion models to predict volatility with precision. By integrating the forecasting results with quantitative investing principles, we constructed a new strategy that achieved better returns in twelve selected American financial stocks. For investors navigating the real stock market, our findings serve as a valuable reference, potentially securing an average annualized return of approximately 5 to 10% for the American financial stocks under scrutiny in our research.

Stocks are the most common financial investment products and attract many investors around the world. However, stock price volatility is usually uncontrollable and unpredictable for the individual investor. This research aims to apply different machine learning models to capture the stock price trends from the perspective of individual investors. We consider six traditional machine learning models for prediction: decision tree, support vector machine, bootstrap aggregating, random forest, adaptive boosting, and categorical boosting. Moreover, we propose a framework that uses regression models to obtain predicted values of different moving average changes and converts them into classification problems to generate final predictive results. With this method, we achieve the best average accuracy of 0.9031 from the 20-day change of moving average based on the support vector machine model. Furthermore, we conduct simulation trading experiments to evaluate the performance of this predictive framework and obtain the highest average annualized rate of return of 29.57%.

Stock price prediction is a core challenge in quantitative finance. While machine learning has advanced the modeling of complex financial time series, existing methods often rely on single-target predictions, underutilize multidimensional market information, and are disconnected from practical trading systems. To address these gaps, this research develops a hybrid machine learning framework for flexible target forecasting and systematic trading of major American technology stocks. The framework integrates Ensemble Models (AdaBoost, Decision Tree, LightGBM, Random Forest, XGBoost) with Fusion Models (Voting, Stacking, Blending) and introduces a Transfer Learning method enhanced by Dynamic Time Warping to facilitate knowledge sharing across assets, improving robustness. Focusing on ten key stocks, we forecast three distinct momentum indicators: next-day Closing Price Difference, Moving Average Difference, and Exponential Moving Average Difference. Empirical results demonstrate that the proposed Transfer Learning approach achieves superior predictive performance and trading simulations confirm that strategies based on these predicted momentum signals generate substantial returns. This research demonstrates that the proposed hybrid machine learning framework can mitigate the high information entropy inherent in financial markets, offering a systematic and practical method for integrating machine learning with quantitative trading.

Option pricing has become a popular topic in the fields of finance and mathematics with the rapid development of stock and option markets. Now, more and more academics, financial companies and investors are attracted to study and do research about it. The theory of option pricing can also be used to price financial instruments with the similar structure to options and contribute to risk control and management. The Black-Scholes model is the basic and famous method applied for different options pricing with modifications and adjustments, and the results can be solved by some traditional numerical methods such as the binomial model, finite difference method, Monte Carlo method and so on. Machine learning has risen recently and begins to replace some complex work in traditional methods with the evolution of computers and computing power. How to use machine learning methods to predict the option price is a problem worthy to be solved. In this research, using the antithetic Monte Carlo method generates the prices of the up-and-out barrier options without rebate based on the Black-Scholes model. The generated dataset is divided into a training set and a test set for support vector regression, random forest, adaptive boosting and artificial neural networks. We compare the fitting and performance of all machine learning methods and find that random forest and artificial neural network methods fit better than others with fewer errors in predictions.