SUMMARY

I am interested in using machine learning to solve complex and challenging problems. I have research experience for over four years, and my research leverages techniques from various fields such as data analysis, natural language processing, computer vision, computational intelligence, and personalization to build powerful models that benefit from diverse and big data.

EDUCATION

• University of Virginia (expected August 2017)

Ph.D. in Systems & Information Engineering

Advisor: Matthew S. Gerber, Assistant Professor

Dissertation: Advanced Crime Prediction Modeling

- Presidential Fellowship in Data Science, Period: 2016 - 2017, Amount: $38,225, Project: Predicting Community-Level Criminal Behaviors by Estimating Human Attitudes from Social Media

• Mississippi State University (2014)

M.S. in Computational Engineering

- Fulbright Scholarship, Period: 2012 - 2014, Amount: $112,820

• Arab International University (2011)

B.S. in Informatics Engineering specializes in Artificial Intelligence

- 1st Rank: ranked the top student

RESEARCH EXPERIENCE

Research Assistant, University of Virginia (May 2014 - Present)

• Crime prediction using localized approaches (August 2015 - Present)

- Developing a new method to perform kernel density estimation (KDE) using ideas from image processing. The proposed method allows for learning of dynamic bandwidths and non-linear kernel functions. Also, the proposed method has a computational complexity that is independent from the number of data points which allows for computing density estimates using large data sets. When tested on real crime data from Chicago, Illinois, the new approach achieved peak gains up to 25% compared with the traditional KDE.

- Implementing area-specific crime prediction models using multi-task learning and hierarchical modelling. The new models improved the prediction performance on 12 out of 17 crime types from Chicago, Illinois.

•Personalized sentiment analysis on Amazon product reviews (April 2015 - August 2015)

- Developing personalized sentiment classification models by applying model adaptation on a global classifier fitted on a separate data set.

- Designing a collaborative learning approach to address the sparseness of the data and allow for learning personalized models for users with a small number of observations.

•Crime prediction using social media (August 2014 - August 2015)

- Modelling individuals' movements and daily routine activities from social media outlets such as Twitter posts and FourSquare venues.

- Improving upon existing crime prediction models by including micro-level movement features estimated using Venue Frequency-Inverse Route Frequency weighting scheme.

•Predicting home location using Twitter (May 2014 - August 2014)

- Designing an algorithm to predict home coordinates of users on Twitter by analyzing their activity diagrams and detecting their sleeping patterns.

- Improving burglaries prediction models by including features about individuals' home occupancy patterns.

TECHNICAL SKILLS

•Programming Languages: C#, JAVA, C++, C, R, Python, SQL(DML, DDL), ASP, JSP, JSF, HTML, CSS, JavaScript, AJAX, XML, XSLT, RDF, OWL

•Databases: MySQL, PostgreSQL, Microsoft SQL, Oracle, SQLite

•IDEs: Microsoft visual studio, Eclipse, Netbeans, RStudio, Spyder

SELECTED PUBLICATIONS

1. M. Al Boni, and M. S. Gerber. Automatic Optimization of Localized Kernel Density Estimation for Hotspot Policing. The 15th IEEE International Conference on Machine Learning and Applications (ICMLA), 2016.

2. M. Al Boni, and M. S. Gerber. Area-Specific Crime Prediction Models. The 15th IEEE International Conference on Machine Learning and Applications (ICMLA), 2016.

3. M. Al Boni, and M. S. Gerber. Predicting Crime with Routine Activity Patterns Inferred from Social Media. International Conference on Systems, Man, and Cybernetics (SMC), 2016.

4. L Gong, M. Al Boni, and H. Wang. Modeling Social Norms Evolution for Personalized Sentiment Classification. The 54th annual meeting of the Association for Computational Linguistics (ACL), 2016.

5. M. Al Boni, K.Q. Zhou, H. Wang, and M. S. Gerber. Model Adaptation for Personalized Opinion Analysis. The 53th annual meeting of the Association for Computational Linguistics (ACL), 2015.

6. M. Al Boni, D.T. Anderson, and R.L. King. Constraints preserving genetic algorithm for learning fuzzy measures with an application to ontology matching. In 2013 Third Annual World Conference on Soft Computing (WCSC 2013), Dec. 2013.

REFERENCES

Matthew S. Gerber, Assistant Professor of Systems & Information Engineering, University of Virginia, acx0x1@r.postjobfree.com, 419-***-****

Hongning Wang, Assistant Professor of Computer Science, University of Virginia, acx0x1@r.postjobfree.com, 434-***-****

Contact this candidate