Tezislər / Theses

 
Figure 2.
Video frame example 

---

THE 3
rd
 INTERNATIONAL SCIENTIFIC CONFERENCES OF STUDENTS AND YOUNG RESEARCHERS 
dedicated to the 99
th
anniversary of the National Leader of Azerbaijan Heydar Aliyev
123
Conclusion 
This research is aimed to detect any traffic accident in real time either 
from surveillance cameras or register cameras by the utilization of 
Convolutional Neural Networks (CNN). Project is in initial stage and has 
ability to predict an accident with 34% accuracy. 
References
1. https://pubmed.ncbi.nlm.nih.gov/31907688/ 
2. https://dl.acm.org/doi/abs/10.1007/s10916-019-1484-1 
3. https://www.sciencedirect.com/science/article/pii/S2405959516300169 
4. https://github.com/lopezbec/Traffic_Accident_Detection 
USING ARTIFICIAL INTELLIGENCE IN AUTOMATION OF 
FINANCIAL RISK MANAGEMENT 
Mahira Asadzade 
Baku Higher Oil School
Baku, Azerbaijan 
[e-mail] 
Supervisor: Ph.D Associate Professor Leyla Muradkhanli
Keywords:
artificial intelligence, machine learning, risk management in finance, deep 
learning, risk analysis, volatiliy forecasting, credit scoring, market risk, claims modelling. 
Financial risk management avoids losses and maximizes profits, and 
therefore is vital to most businesses. The ability of machine learning models 
to analyze large amounts of both structured and unstructured data can 
improve analytical capabilities in risk management and compliance, allowing 
risk managers in financial institutions to identify risks in an effective and 
timely manner, make more informed decisions, and make banking less risky. 
Based on the source of risk factors, it is conventional to classify risk into four 
high-level categories: 
Market risk
, 
Credit risk
, 
Insurance and demographic 
risk
, and finally, 
Operating risk
[1]. 
Market risk is the risk arising from changes in financial indicators, such 
as the exchange rate, interest rate, inflation, and so on. One of the key 
measures related to market risk is 
volatility
which reflects the uncertainty of 
future asset prices in the financial market.
Gen- eralised Autoregressive 
Conditional Heteroskedasticity (GARCH) model is an autoregressive moving 
average model for conditional variance with p number of lagged squared 
returns and q number of lagged conditional variance, so it is widely used by 
practitioners for modelling time-variant volatility of financial time series [2].
Data-driven models, namely Support Vector Regression and Neural Networks 
were applied to significantly improve forecast performance and the results 
are shown below.

---