Tọa đàm Toán ứng dụng và Khoa học dữ liệu Tháng 11 - 2023

Mục tiêu:
-      Xây dựng diễn đàn để các nhà nghiên cứu trình bày, thảo luận những kết quả nghiên cứu mới trong ngành Toán nói chung, các hướng nghiên cứu về khoa học dữ liệu, và các ứng dụng của toán và khoa học dữ liệu trong kinh doanh.
-      Thúc đẩy việc giao lưu, trao đổi học thuật nhằm kết nối cộng đồng nghiên cứu đồng thời góp phần nâng cao chất lượng hoạt động của giảng viên, nghiên cứu viên của Trường.

Thời gian: 7g30-11g05, ngày11/11/2023

Địa điểm: Hội trường Lầu 10 Nhà B, Cơ sở 56 Hoàng Diệu 2, Thủ Đức, Hồ Chí Minh

Tóm tắt các chủ đề & báo cáo viên trong toạ đàm

GS. TS. Lê Văn Thuyết

College of Education, Hue University

Title: Coding theory and Frobenius rings

Abstract: In this talk, we will give the basic concepts in coding theory and finite ring theory. Especially, we present some results about (quasi)-Frobenius rings. In coding theory, now we also change a finite field of code alphabet set to a commutative finite ring. We will answer the question: Why should it be Frobenius? 

TS. Nguyễn Đức Phương

Industrial University of HCM City

Title: On the Stochastic Elliptic equation involving Fractional derivative.

Abstract: This study is focused on finding the solution to the initial value problem for the fractional elliptic equations driven by the Wiener process. At first, this research demonstrates that our forward problem is ill-posed in the sense of Hadamard. For this ill-posed problem, the truncation method is used to construct a regularized solution. Under prior assumptions for the exact solution, the convergence rate is obtained.

PGS. TS. Đinh Thanh Duc

Quy Nhon University

Title: Some problems on fractional partial differential equations and their applications

Abstract: In this report, we provide an overview of the history of fractional calculus problems and present novel research directions regarding partial differential equations with fractional orders in specific types of spaces. Additionally, we introduce new findings related to various multi-term Caputo and Riemann-Liouville fractional integro-differential equations within the Wiener space of functions.

Simultaneously, we will discuss the applications of fractional calculus in the context of economic modeling. Fractional calculus has been applied in various economic modeling and analysis scenarios. Here are some ways in which fractional calculus can be used in the field of economics.

ThS. Bui Thi Thien My

HCM University of Banking

Title: An interpretable decision tree ensemble model for imbalanced credit scoring datasets

Abstract: Credit scoring is a typical example of imbalanced classification, which poses a challenge to conventional machine learning algorithms and statistical classifiers when attempting to accurately predict outcomes for defaulting customers. In this paper, we propose a credit scoring classifier called Decision Tree Ensemble model (DTE). This model effectively addresses the challenge of imbalanced data and identifies significant features that influence the likelihood of credit status. An experiment demonstrates that DTE exhibits superior performance metrics in comparison to well-known based-tree ensemble classifiers such as Bagging, Random Forest, and AdaBoost, particularly when integrated with resampling techniques for handling imbalanced data. 

TS. Vo Duc Vinh

HCM University of Banking

Title: Mining User Preferences on Social Networking Platforms and Its Applications in Online Marketing

Abstract: With the increase in use of social networks, users on these platforms usually share their demographic information, thoughts and interests via short texts such as posts, status, or reactions. Capturing user preferences from these kinds of data brings a lot of benefits for business activities such as item recommendation in e-commerce systems, job offering in labour markets, or advertising to potential customers in online marketing.

However, seeking an efficient solution for the problem of mining user preferences on these platforms is not trivial but a challenging task. Researchers working on this problem usually face following challenges: (1) The data sparsity and cold-start issue existing in user texts; (2) user preferences dynamically change over time; (3) social networking users usually create lots of short documents. The consecutive documents are often not very closely related to each other. This causes difficulties for inferring user preferences; (4) data may come in different formats (e.g., images, texts, or reactions) from multiple sources (e.g., one user may simultaneously have multiple accounts across networking platforms).

This talk will present an approach for mining user preferences in two scenarios, the static and the dynamic. The static framework is useful for mining user demographic information while the dynamic framework is valuable in capturing the change of user preferences over time. Essentially, these proposals combine advancements in machine learning, deep learning, and evidential theory to infer user preferences. Two practical datasets, the users’ data from Facebook and Twitter, are used to conduct experiments on the proposed frameworks.

To demonstrate the applicable aspect of proposed frameworks in business scenarios, we introduce a method for visualizing the fluctuation of user preferences on various topics over time (i.e., economics, politics, entertainments, or education). This visualization may reveal significant insights that are useful for many practical applications such as item recommendation in e-commerce systems, job offering in labour markets, or advertising in online marketing.

Bộ môn Toán kinh tế kính mời các thầy/cô và các bạn sinh viên từ các Khoa/Bộ môn đến tham dự.

Trân trọng