Gallery

Year 2022

CMS Research Sharing Seminar 2022

Date: 09/12/2022 (Friday)

Time: 9.00a.m to 4.30p.m

Platform: Microsoft Team

Organizer: Centre of Mathematical Science, CMS

Research Talk by Invited Speaker from Universiti Putra Malaysia 

Research Title: Sparse Optimization: Simple is Better

Date

25 March 2022 (Friday)

Time

2.00pm – 3.00pm

Platform

Microsoft Teams

Speaker

Prof. Dr. Leong Wah June

Department of Mathematics, Faculty of Sciences

Universiti Putra Malaysia

Organizer/ 


Centre for Mathematical Sciences (CMS)

Abstract: In this talk, we first introduce and motivate the application of sparse optimization in wide variety of fundamental applications. A general proximal alternating linearized (PAL) method for solving a broad class of sparse optimization problems involving L0-norm is proposed. Building on the Kurdy-Lojasiewicz property, a brief convergence analysis is established. We then demonstrate the applicability of the PAL method in solving some real-life problems such as those in sparse optimal control of large-scale interconnected networks, portfolio optimization and image restoration.

Research Talk by Invited Speaker from Universiti Teknologi MARA (UiTM) 

Research Title: The classification Problem of Some Classes of Finite-Dimensional Algebras

Date

24 June 2022 (Friday)

Time

3.00pm – 4.00pm

Platform

Microsoft Teams

Speaker

Prof. Isamiddin Rakhimov

Faculty of Computer and Mathematical Sciences

Universiti Teknologi MARA (UiTM)

Organizer/ 


Centre for Mathematical Sciences (CMS)

Abstract: The talk devoted to the classification problem of finite-dimensional algebras. We review the general strategy applied so far. Then the classification problem of Lie algebras and certain classes of algebras closely related to the class of Lie algebras will be focused on. These classes of algebras wee introduced by J.L. Loday in 1994. We show links between these classes of algebras and provide the latest results on their classification problem in low dimensions.

Research Talk by Invited Speaker from Universiti Putra Malaysia (UPM) 

Research Title: Comparison of Power Option Prices of Asymmetric Barrier Type

Date

29 September 2022 (Thursday)

Time

3.00pm – 4.00pm

Platform

Microsoft Teams

Speaker

Prof Madya Dr. Siti Nur Iqmal Binti Ibrahim

Department of Mathematics and Statistics, Faculty of Science

Universiti Putra Malaysia

Organizer/ 


Centre for Mathematical Sciences (CMS)

Abstract: The basis of options has been the vanilla options, and many have studied extensions of this option to construct other options, called the exotic options, to attract investors. One such option is the power options. With its higher leverage with every change in the underlying asset, power options are deemed to be more profitable than the vanilla options.

In this work, we study pricing models for European-style power call options and a modified version of this option. First, we present the fast Fourier transform (FFT) option pricing model for power call options, and also apply the Monte Carlo simulation (MCS) to compare the efficiency between these two pricing models. Then, we present a pricing formula for power call options that includes a barrier, which is called the power down-barrier options. We also show that there exists a power put-power call parity relationship, and a transformation between the underlying asset and the power contract. This study is within the Black0Scholes environment.

Finally, we run numerical experiments where we first compare the FFT and MCS for pricing the power call options, and then we compare the prices between power call options and power down-barrier options. 

Research Talk by Invited Speaker from The Hong Kong Polytechnic University

Research Title: On the Optimization of Acoustric Beamformer Systems and Configurations

Date

28 July 2022 (Thursday)

Time

3.00pm – 4.00pm

Platform

Microsoft Teams

Speaker

Prof. Yiu Ka-Fai, Cedric

Department of Applied Mathematics

The Hong Kong Polytechnic University

Organizer/ 


Centre for Mathematical Sciences (CMS)

Abstract: For signal processing, multi-sensors are often deployed in order to enhance signal quality in various ways. In general, beamforming technique can be applied via an array of sensors to enhance the required signal via spatial filtering. Designing better beamformers for different scenarios is essential when the sensor distribution and the environment has changed. Optimization plays an important role in the improvement process. In this talk, we will discuss some advances in the optimal design of multi-dimensional broadband beamforming system. We review on various approaches and discuss some of the performance issues. Different optimization models will be considered. In addition to optimizing filter coefficients, we found that the geometric configuration of the array is important for the accuracy of the designs. In view of this, microphone locations can be optimized together with the filter coefficients and the overall problem is formulated as a non-convex optimization problem. When wireless sensors are operated, it is possible to design the beamformer in a distribution manner. Furthermore, it is possible to employ only a subset of the sensors to satisfy a specified performance requirement. In this way, the complexity of the beamformers can be greatly reduced. We will illustrate by several examples the proposed methods.

Research Talk by Invited Speaker from Heriot-Watt University, Malaysia 

Research Title: A First-Order Spatial Integer-Valued Autoregressive Sinar (1,1) Model

Date

26 August 2022 (Friday)

Time

2.30pm – 3.30pm

Platform

Microsoft Teams

Speaker

Dr. Mahendran Shitan

School of Mathematical and Computer Science

Heriot-Watt University, Malaysia

Organizer/ 


Centre for Mathematical Sciences (CMS)

Abstract: In this paper, we extend the idea of Gegenbauer process in the spatial domain by introducing a more general parameter and call this model as Spatial Gegenbauer Autoregressive (SGAR (1,1)) model. The spectral density and autocovariance functions of the model are introduced. The Yajima estimators of the Gegenbauer parameters, the log-periodogram regression estimators of the memory parameters and the Whittle's estimators of all parameters are discussed. The performance of these estimators is evaluated through a simulation study.

Research Talk by Invited Speaker from International Islamic University Malaysia (IIUM)

Research Title: Listening to Einstein's Universe

Date

31 October 2022 (Monday)

Time

3.00pm – 4.00pm

Platform

Microsoft Teams

Speaker

Ts. Dr. Hafizah Noor Bt Isa

Department of Physics, Kulliyyah of Science

International Islamic University Malaysia (IIUM)

Organizer/ 


Centre for Mathematical Sciences (CMS)

Abstract: Einstein's theory of general relativity predicted the existence of gravitational waves, but it was only in 2015 that they were finally detected by scientists working at LIGO laboratories. This discovery has been heralded as one of this century's greatest scientific breakthroughs and may have implications for our understanding not just dark matter or energy distributed throughout space-time- but time itself.

Research Talk by Invited Speaker from Taylor's University

Research Title: Multidimensional Functional Relationship Model with Multicollinearity

Date

29 April 2022 (Friday)

Time

2.00pm – 3.00pm

Platform

Microsoft Teams

Speaker

Associate Professor Dr. Chang Yun Fah

Actuarial Studies Program, School of Accounting and Finance, Faculty of Business and Law

Taylor's University

Organizer/ 


Centre for Mathematical Sciences (CMS)

Abstract: The functional relationship model (also called measurement error model or error in variables model) is a model extended from the classical linear regression where errors are assumed in both the X and Y variables which has been explored since the 19th century by Adcock (1878). Many further studies were then conducted to p-dimension of X and Y with single slope and uncorrelated variables. In this talk, we discuss a generalized version of Chang et al. (2010) works by allowing multicollinearity among the dimensions under study. The fundamental derivations on the model which includes multicollinearity to obtain the close form equations of the parameters and the properties of the estimated parameters will be discussed. At the end of the talk, simulation results and numerical example will also be presented.

Year 2021

Virtual Visit of External Examiner 

Date: 6th & 7th December 2021 

Platform: Microsoft Teams


Programme: Bachelor of Science (Honours) Applied Mathematics with Computing 

External Examiner: Prof. Dr. Huang Huang-Nan (Tunghai University, Taiwan)


Programme: Bachelor of Science (Honours) Financial Mathematics

 External Examiner: Prof. Dr. Yang Hailiang (The University of Hong Kong)

Opening Meeting

Meeting with Staff

Research Talk by Prof. Dr. Yang Hailiang

Optimal Insurance Strategies: A Hybrid Deep Learning Markov Chain Approximation Approach

Date

6 December 2021 (Monday)

Time

2.00pm – 3.00pm

Platform

Microsoft Teams

Speaker

Prof. Dr. Yang Hailiang

Department of Statistics and Actuarial Science

The University of Hong Kong

Hong Kong

Organizer/ 

Co-organizer

Department of Mathematical and Actuarial Sciences (DMAS)

Centre for Mathematical Sciences (CMS)

Abstract:

This paper studies deep learning approaches to find optimal reinsurance and dividend strategies for insurance companies. Due to the randomness of the financial ruin time to terminate the control processes, a Markov chain approximation-based iterative deep learning algorithm is developed to study this type of infinite-horizon optimal control problems. The optimal controls are approximated as deep neural networks in both cases of regular and singular types of dividend strategies. The framework of Markov chain approximation plays a key role in building the iterative equations and initialization of the algorithm. We implement this self-learning approach to approximate the optimal strategies and compare the learning results with existing analytical solutions. Satisfactory computation efficiency and accuracy are achieved as presented in numerical examples.


Thank you for the informative sharing!


Meeting with Students (AM)

Meeting with Students (FM)

Exit Meeting

Virtual Visit of External Examiner for Bachelor of Science (Honours) Actuarial Science


Prof. Dr. Yam Sheung Chi Phillip

(The Chinese University of Hong Kong)

Date: 29th & 30th November 2021

Platform: Zoom

Opening Meeting

Meeting with Students

Meeting with Staff

Exit Meeting


Research Talk by Invited Speaker from Universiti Teknologi PETRONAS

Multi-EigenSpot: A Novel Eigenspace-Based Method for Detecting Multiple Space-Time Disease Clusters

Date

21 October 2021 (Thursday)

Time

10.00am – 11.00am

Platform

Microsoft Teams

Speaker


Assoc. Prof. Dr Hanita binti Daud

Fundamental and Applied Sciences Department

Universiti Teknologi PETRONAS

Perak, Malaysia

Organizer/ 

Co-organizer

Department of Mathematical and Actuarial Sciences (DMAS)

Centre for Mathematical Sciences (CMS)

Abstract:

Detecting the potential space-time disease clusters is necessary for conducting surveillance and implementing disease prevention policies. In disease surveillance, space-time cluster detection aims to detect the sub-regions in the spatiotemporal space, where the observed disease count is higher than what is to be expected if no risk factor exists. The state-of-the-art method for this problem is the space-time scan statistic (SaTScan). This method is based on the Maximum Likelihood Estimation (MLE) which put some constraints on the distribution of the data such as Poisson or Gaussian counts that are valid only for the laboratory data and not necessarily valid for the non-traditional data sources. Addressing this problem, an Eigenspace-based method called an EigenSpot has been recently proposed as a nonparametric solution for space-time cluster detection. However, the main problem with the EigenSpot method is that it can detect a single cluster only and cannot be adapted for detecting multiple clusters. This is an important limitation, since multiple hotspots may have occurred in the study area, sometimes at the same level of importance. Addressing this issue, this study proposes an extension of the EigenSpot method, called a Multi-EigenSpot that is able to handle multiple hotspots by iteratively removing previously detected hotspots and re-running the algorithm until no more hotspots are found. The proposed method uses the expected disease counts as the baseline information instead of the population counts and hence allows for removing the previously detected cluster by replacing the observed counts by the respective expected counts. The proposed method approximates multiple clusters automatically through a visualization tool. In addition, the proposed method can be adapted for detecting clusters with unknown population counts which is often the case in the least developed countries. A comprehensive experimental evaluation, both on simulated and real-world datasets reveals that the proposed method not only provides a nonparametric solution to multiple clusters detection problem but also characterizes relatively higher efficiency than the state-of-the-art methods.

 

Thank you for the wonderful sharing!



Research Talk by Invited Speaker from Multimedia University Melaka, Malaysia

Swarm Intelligence Mimicking to Solve Complex Problems

Date

 10 September 2021 (Friday)

Time

10.00am– 11.00am

Platform

Microsoft Teams

Speaker

Dr Nor Azlina Binti Ab Aziz

Faculty of Engineering and Technology

Multimedia University Melaka, Malaysia

Organizer/ 

Co-organizer

Department of Mathematical and Actuarial Sciences (DMAS)

Centre for Mathematical Sciences (CMS)

Abstract:

Swarm intelligence (SI) belongs to the family of computational intelligence algorithms. The algorithms from SI family are mostly inspired from nature and work based on the concept of collaboration and information sharing. Among the popular SI algorithms are Particle Swarm Optimization (PSO), Ant Colony Optimization (ACO), Artificial Bee Colony (ABC), and many others. These algorithms share the same framework but differ from each other on how the agents look for a solution which is dependent on their source of inspiration. During this session, the broad concept of SI will be discussed, followed by some of the popular SI algorithms and an example of application for solving a real-world problem. This session aims to attract the interest of new researchers in this field and as a channel of discussion among existing SI’s researchers.

 

Thank you for the informative sharing!



Research Talk by Invited Speaker from Universiti Tun Hussein Onn Malaysia

Curve Fitting with First-Order Linear Differential Equation

Date

 20 August 2021 (Friday)

Time

10.00am– 11.00am

Platform

Microsoft Teams

Speaker

Dr Kek Sie Long

PhD, CQRM

Department of Mathematics and Statistics

Universiti Tun Hussein Onn Malaysia

Pagoh Campus, Muar, Johor, Malaysia

Organizer/ 

Co-organizer

Department of Mathematical and Actuarial Sciences (DMAS)

Centre for Mathematical Sciences (CMS)

Abstract:

In this talk, the curve fitting, which is handled by using the first-order linear differential equation, is discussed. For this purpose, the coefficient of proportionality is determined from the set of data points. For the constant coefficient, the measures of the central tendency, which are mean, median, and mode, are calculated. By applying these constant coefficients and the varying coefficients, the first-order linear differential equation exists and solvable. On this basis, the Euler method is applied to approximate the solution of the differential equations. For illustration, some sets of data points and their curves are observed, while the best fitting of these curves is demonstrated by using the method proposed. The results show that the performance efficiency of the method proposed is satisfied within a given tolerance. In conclusion, applying the linear differential equation to fit the curve of data points provides an effective alternative approach for the curve fitting problem.

 

Thank you for the educational sharing!



Research Talk by Invited Speaker from Universiti Sains Malaysia

Tight Bounds on the Burning Numbers of Path Forests and Spiders

Date

 13 August 2021 (Friday)

Time

10.00am– 11.00am

Platform

Microsoft Teams

Speaker

Dr Teh Wen Chean

School of Mathematical Sciences

Universiti Sains Malaysia

Pulau Pinang, Malaysia

Organizer/ 

Co-organizer

Department of Mathematical and Actuarial Sciences (DMAS)

Centre for Mathematical Sciences (CMS)

Abstract:

In 2016, Bonato, Janssen, and Roshanbin introduced graph burning as a discrete process that models the spread of social contagion. Although the burning process is a simple algorithm, the problem of determining the least number of rounds needed to completely burn a graph is NP-complete even for elementary graph structures like starlike trees, also called spiders. An open burning number conjecture states that every connected graph of order m2 can be burned in at most m rounds. Attempts to prove the conjecture have resulted in various upper bounds for the burning number and verification of the conjecture for various classes of graphs, including spiders. In this seminar, we will present a detailed outline about our contribution towards the burning number conjecture, which results from a collaborative work with Ta Sheng Tan. Together, we found a tight upper bound on the order of a spider for it to be burned within a given number of rounds. Our result showed that the tight bound depends on the structure of the spiders, namely the number of arms. Additionally, an analogous corresponding tight upper bound for path forests was obtained, thus completing a previously known partial result. This suggests new perspective towards the burning number conjecture via characteristics of graphs. We will end the presentation by briefly mentioning the speaker's ongoing work as appropriate. This presentation is intended to facilitate readers interested in our contribution and as an invitation for potential collaboration into this work.


 

Thank you for the interesting sharing!



Research Talk by Invited Speaker from Universiti Malaya

Modelling and forecasting stock volatility and return: A new approach based on quantile Rogers-Satchell volatility measure with asymmetric bilinear CARR model

Date

 6 August 2021 (Friday)

Time

11.00am– 12.00pm

Platform

Microsoft Teams

Speaker

Assoc Prof Dr Ng Kok Haur

Institute of Mathematical Sciences

Faculty of Science

Universiti Malaya

Kuala Lumpur, Malaysia

Organizer/ 

Co-organizer

Department of Mathematical and Actuarial Sciences (DMAS)

Centre for Mathematical Sciences (CMS)

Abstract:

Volatility of asset prices in financial market is not directly observable. Return-based models have been proposed to estimate the volatility using daily closing prices. Recently, many new range-based volatility measures were proposed to estimate the volatility directly. This study proposes quantile Rogers-Satchell (QRS) measure to ensure robustness to extreme prices. We add an efficient term to correct the downward bias of Rogers-Satchell (RS) measure and provide scaling factors for different interquantile range levels to ensure unbiasedness of QRS. Simulation studies confirm the efficiency of QRS measure relative to the intraday squared returns and RS measures in the presence of extreme prices. To smooth out the noises, QRS measures are fitted to the conditional autoregressive range model with different asymmetric mean functions. By comparing to two realised volatility measures as proxies for the unobserved true volatility, result from Standard and Poor 500 index shows that QRS volatility estimates using asymmetric bilinear mean function provide the best in-sample model fit based on two robust loss functions. These fitted volatilities are then incorporated into return models to capture the heteroskedasticity of returns. Model with a constant mean, Student-t errors and QRS estimates gives the best-in-sample fit. Different value-at-risk (VaR) and conditional VaR forecasts based on the best return model are also provided and tested.


 

Thank you for the wonderful sharing!



Research Talk by Invited Speaker from Universiti Malaysia Terrengganu (UMT)

Ordered Weighted Averaging Operators: Modelling the Majority Concept based on Cardinality

Date

 30 July 2021 (Friday)

Time

11.00am– 12.00pm

Platform

Microsoft Teams

Speaker

Dr. Binyamin Yusoff

Mathematical Sciences Field

Faculty of Ocean Engineering Technology and Informatics

Universiti Malaysia Terrenganu (UMT)

Terrengganu, Malaysia

Organizer/ 

Co-organizer

Department of Mathematical and Actuarial Sciences (DMAS)

Centre for Mathematical Sciences (CMS)

Abstract:

Aggregation of several input values into a single output value is crucial in numerous mathematical models. The problems of aggregation are very broad and heterogeneous. This talk will particularly focus on the ordered weighted averaging (OWA) operators, a specific topic of aggregation under the finite number of real inputs. The OWA operators are a parameterized family of mean operators. Recent developments of OWA, especially under the cardinality-dependent aggregation operators, or the so-called majority-additive OWA, will be presented. The MA-OWA operator generalizes the simple arithmetic mean, and it is known as the arithmetic mean of arithmetic means. Moreover, it can control the effect of extreme values in arithmetic mean that may cause biased results. The other variants of this aggregation operator, namely selective majority additive (SMA-OWA), selective aggregated majority (SAM-OWA) and weighted SAM-OWA with additional characteristic of cardinality relevance factor (CRF) will also be highlighted. Finally, the application of these aggregation operators in group decision-making problem will be presented to conclude this talk.

 

Thank you for the enlighten sharing!



Research Talk by Invited Speaker from Universiti Teknologi MARA

Big Data Analytics: Tools, Techniques and Applications

Date

 16 July 2021 (Friday)

Time

11.00am– 12.00pm

Platform

Microsoft Teams

Speaker

Prof Dr. Yap Bee Wah

Institute for Big Data Analytics and Artificial Intelligence (IBDAAI)

& Center of Statistical and Decision Science Studies

Faculty of Computer and Mathematical Sciences (FSKM)

Universiti Teknologi MARA, Shah Alam, Selangor, Malaysia

Organizer/ 

Co-organizer

Department of Mathematical and Actuarial Sciences (DMAS)

Centre for Mathematical Sciences (CMS)

Abstract:

Big Data Analytics involves discovering useful and meaningful insights from structured and unstructured data.  Data collection and storage is growing exponentially through the use of advanced digital technology, IoT and drones. Machine Learning and Deep Learning are useful algorithm for Big Data Analytics. This seminar covers the Big Data Analytics tools, techniques and applications in various domains. We present the Big Data Analytics framework, the popular data mining tools and machine learning algorithms for prediction and classification problem. The aim of this seminar is to share the potential applications of big data analytics for industry and academic research.

Thank you for the informative sharing!