Emilio Sagre, Gustavo Dutra, Niha Siddikha Sachin, Steven Whang 
Data Analytics, Langara College

Creating Accurate Predictions in the Stock Market

Time & Date:  5 p.m., Wednesday, December 14th, 2022
Location: Room HS301 and via Zoom (see an email and registration information)
Registration is open now: https://events.vtools.ieee.org/m/337514

Abstract:

Creating accurate predictions in the stock market has always been a great challenge in finance. With the rise of machine learning as the next level in the forecasting area, this research compares four machine learning models and their accuracy for forecasting three well-known stocks traded in the NYSE in the short term over the period from March 2020 to May 2022. We deploy, develop, and hypertune XGBoost, Random Forest, Multi-layer Perceptron, and Support Vector Regression models and report the models that produce the highest accuracies from our evaluation metrics: RMSE, MAPE, and MPE.

Speakers Bio:

• Emilio Sagre is a licensed Civil Engineer in the Philippines. He is a student at Langara College finishing his Post-Degree Diploma (PDD) in Data Analytics. His passion for machine learning and its practical applications led him to enter this field.
• Gustavo Dutra has a Bachelor of Engineering and works in the Tech Industry, helping companies improve their data management and operations. Being passionate about business data, he’s also finishing his PDD in Data Analytics at Langara College.
• Niha Siddikha Sachin In the 5 years that Niha spent crunching numbers for the e-commerce and retail industries as a lead and a product analyst in a trillion-dollar and a Fortune 500 company, respectively, she has solved problems for various categories, created reports to track PnL and maintained servers for the banking and finance domain. Having a Bachelor of Engineering degree in IT, she became interested in data interpretation and manipulation, where data alter the perception and when the perception alters output. Niha is currently finishing her PDD in Data Analytics at Langara college.
• Steven Whang has a Bachelor’s Degree in Financial Economics. He is finishing a PDD in Data Analytics at Langara College. He aspires to be working in the data science field in the future.

For further information, please contact: Youry Khmelevsky (email: Youry at IEEE.org) and subscribe for the news at https://listserv.ieee.org/cgi-bin/wa?SUBED1=okanagan&A=1)
Pizza and Refreshments will be provided

Evan MacKinnon
Computer Science Department
Okanagan College

and

Dakota Joiner
Computer Science Department
Okanagan College

An approach towards generating k-leaf powers for phylogenetic tree construction

Time & Date:  4 p.m., Wednesday, April 27, 2022
Location: Room E102 and via Zoom (see email and registration information), Computer Science Department, Okanagan College
Registration is open now: https://events.vtools.ieee.org/m/312801

Abstract:

Phylogeny is the evolutionary history of a species or group of organisms. Evolutionary trees can be analogized to graph trees, thus determination of these structures aids in inferring the evolutionary history of groups of organisms, extant and extinct. K-leaf power graphs enhance the ability of researchers to map paralogous and xenologous speciation events in what is a considerably difficult area to correctly predict past relationships.

Forbidden subgraph characterization is a method by which to characterize a graph class with a set of graphs that do not belong to that class. Identification of minimal forbidden induced
subgraphs is one method of characterizing the k-leaf powers of graph trees. A tree is a k-leaf power if, and only if, the leaves are connected by at most distance k. Structures for 2-leaf, 3-leaf, and 4-leaf powers are well understood, however, there does not exist a published list of forbidden subgraph leaf powers for values of k ≥ 4. In service of cladistics, k-leaf powers are frequently edited by adding or removing nodes and edges to the closest “proper” representation of a pairwise comparison of groups of organisms.

We demonstrate a deterministic, reductionist approach to generating 4-leaf, 5-leaf, and 6-leaf powers using Python, the graph library Networkx, and the Nauty suite of graph generation and labelling programs. The list of non-k-leaf powers in this range has not been proven finite, so this approach does not cover all possible structures should the list be infinite.

Speakers Bio:

• Evan MacKinnon is currently studying computer science at Okanagan College and will receive his bachelor’s degree in June 2022.
• Dakota Joiner graduated with a Bachelor of Science in Chemistry and a Bachelor of Science in Medical Biochemistry from the University of British Columbia in 2014. He is currently studying computer science and data science at Okanagan College.

For further information please contact: Youry Khmelevsky (email: Youry at IEEE.org) and subscribe for the news at https://listserv.ieee.org/cgi-bin/wa?SUBED1=okanagan&A=1)
Pizza and Refreshments will be provided

Albert Wong, Ph.D.
Math Department, Langara College
IEEE Member
and

Youry Khmelevsky, Ph.D.
Computer Science Department
Okanagan College

Algorithmic Trading and Short-term Forecast for Financial Time Series with Machine Learning Models

Time & Date:  4 p.m., Tuesday, March 29, 2022
Location: Room L318 (up to 16 peoples), Langara College, 100 West 49th Avenue, Vancouver B.C., Canada V5Y 2Z6 and Online (Via email and Registration below)
Registration is open now: https://events.vtools.ieee.org/m/309786

Abstract: A profitable algorithmic stock trading algorithm benefits from a forecasting system that can produce accurate short-term forecasts. Based on this premise, we proposed to build up our previous experience in building short-term forecasting models using machine learning models. The research project aims to develop effective algorithmic trading algorithms based on accurate short-term forecasts for financial time series using machine learning models. The project focuses on three research activities: (1) building a data warehouse containing the targeted financial time series and other time series considered valuable as input to the machine learning models. Data acquisition, processing, and staging routines that are required to feed the data warehouse dynamically are evaluated and developed. A data visualization and business intelligence layer is also built on top of the data warehouse; (2) a short-term forecasting model based on machine learning is developed using the various time-series data from the data warehouse. Machine learning algorithms such as neural network, random forecast, support vector regression, XGBoost, and long short-term memory (LSTM) are evaluated in conjunction with several performance criteria to identify the most accurate model from a short-term trading perspective; (3) an automated evaluation system is developing to assess the effectiveness of existing algorithmic short-term trading strategies. Novel algorithms are also developed and evaluated using the evaluation system and the short-term forecast machine learning model.

The current research project was motivated by 3 NSERC funded projects:

• May 2021 – May 2023. NSERC CCI ARD2: Utilities, Smart metering, Data warehouse, Natural Language Processing, Speech-to-text. Industrial Client: Harris SmartWorks.
• June 2020 – March 2021. A novel approach to COVID-19 Impact Analysis and Reporting for Utilities. Host institution: Langara College. Industrial Client: Harris SmartWorks.
• 2016 – 2019. NSERC CCI ARD2 (477506-14): GPN-Perf2: Game private networks and game servers performance optimization. Industrial Client: WTFast.

Speakers Bio:

• Albert Wong, Ph.D., teaches at the Post-degree Diploma Program in Data Analytics at Langara (the “DA Program”). Drawing on experience and skills developed over a long career in the field, Dr. Wong sets the technical direction for projects proposed within several applied research applications and liaise closely with the partners’ senior leadership teams to ensure that the projects align with each firm’s strategic and financial imperatives. Dr. Wong’s academic expertise includes sampling, multivariate statistical analysis, and ML. He has spent decades working and consulting with organizations of various sizes on the use of statistics, DA, and information technology to solve strategic and/or tactical problems.
• Youry Khmelevsky received his Ph.D. degree in computer science. His current research interests include enterprise-wide DBMS systems, database warehousing; data mining; software engineering; cloud and high-performance computing; enterprise-wide information systems, no programming paradigm and blind computing. Dr. Khmelevsky had served as a postdoctoral fellow at Harvard University; was a Visiting Scientist in the Computer Science and Artificial Intelligence Laboratory (CSAIL), Massachusetts Institute of Technology (MIT); was an Invited Researcher at Database Management and Machine Learning Department, Sorbonne University, Paris, France; held engineering and R&D positions in Industry in Europe and North America for about 15 years, including at Alberta Energy, Government of Alberta, Canada.

For further information please contact: Youry Khmelevsky (email: Youry at IEEE.org) and subscribe for the news at https://listserv.ieee.org/cgi-bin/wa?SUBED1=okanagan&A=1)
Refreshments will be provided