“Narine is staggeringly brilliant and naturally collaborative. It is a pleasure to work with her, both in her capacity as the broad-ranging technical linchpin for Faraday's predictive modeling, but also as an endlessly-inquisitive and patiently-helpful coworker.”
About
Dr. Hall is an educator, expert in machine learning, data-driven entrepreneur, author and…
Contributions
-
What do you do if stakeholders doubt your data mining findings?
I think one of the most important factors in the data mining is really spending time to understand the domain in which data is collected, the main objectives of the stakeholders and how they define success. This process allows to create trust through dialogue and also use know facts in that space to validate any data patterns or predictions. For example, once I used data mining models for a financial landing company and our clustering algorithm found very distinctly separated group of customers, which we later found were actually related to a fraud that they were aware of. Since then they trusted the clustering algorithm to separate their customers into groups based on their similarities in spending patterns.
-
What do you do if stakeholders doubt your data mining findings?
I think educating stakeholders can go a long way when creating shared understanding of success from predictive models. It is equally important to understand the main objectives of the stakeholders and what they really care about or know about the problem, as this can help to really guide the data mining process whether it means understanding better the domain knowledge that is required for accurate identification of the right features, biases in data or temporal patterns to watch out for or use the same shared understanding to contextualize the outcomes for the shareholders and earn their trust by showing that you understand enough about their domain to create credible models.
-
What is the best way to partition data in data engineering for Data Mining?
In my experience one of the most critical parts of data partitioning is to understand if we have enough data for proper training, testing and validation and if the answer is no, how we can accurately generate synthetic data to do this without compromising the quality of data. A great data pre processing step can help with this.
Activity
-
Great to see NU on the list! https://lnkd.in/dQi-UcG4
Great to see NU on the list! https://lnkd.in/dQi-UcG4
Liked by Narine Hall, Ph.D.
-
Skip the branded merch and go for fun statement accessories! 🫣🙃 Mobile case of the expert sales coach Fares Měchura - PARTNER at Saleshero at the…
Skip the branded merch and go for fun statement accessories! 🫣🙃 Mobile case of the expert sales coach Fares Měchura - PARTNER at Saleshero at the…
Liked by Narine Hall, Ph.D.
-
See who you can spot from these pictures. I see Stephen Wolfram and lots of other great speakers! Perusall
See who you can spot from these pictures. I see Stephen Wolfram and lots of other great speakers! Perusall
Shared by Narine Hall, Ph.D.
Experience
Education
Licenses & Certifications
Publications
-
Team Learning for Healthcare Quality Improvement
IEEE Access
In organized healthcare quality improvement collaboratives (QICs), teams of practitioners from different hospitals exchange information on clinical practices with the aim of improving health outcomes at their own institutions. However, what works in one hospital may not work in others with different local contexts because of nonlinear interactions among various demographics, treatments, and practices. In previous studies of collaborations where the goal is a collective problem solving, teams of…
In organized healthcare quality improvement collaboratives (QICs), teams of practitioners from different hospitals exchange information on clinical practices with the aim of improving health outcomes at their own institutions. However, what works in one hospital may not work in others with different local contexts because of nonlinear interactions among various demographics, treatments, and practices. In previous studies of collaborations where the goal is a collective problem solving, teams of diverse individuals have been shown to outperform teams of similar individuals. However, when the purpose of collaboration is knowledge diffusion in complex environments, it is not clear whether team diversity will help or hinder effective learning. In this paper, we first use an agent-based model of QICs to show that teams comprising similar individuals outperform those with more diverse individuals under nearly all conditions, and that this advantage increases with the complexity of the landscape and level of noise in assessing performance. Examination of data from a network of real hospitals provides encouraging evidence of a high degree of similarity in clinical practices, especially within teams of hospitals engaging in QIC teams. However, our model also suggests that groups of similar hospitals could benefit from larger teams and more open sharing of details on clinical outcomes than is currently the norm. To facilitate this, we propose a secure virtual collaboration system that would allow hospitals to efficiently identify potentially better practices in use at other institutions similar to theirs without any institutions having to sacrifice the privacy of their own data. Our results may also have implications for other types of data-driven diffusive learning such as in personalized medicine and evolutionary search in noisy, complex combinatorial optimization problems.
Other authors -
Data-Driven Cluster Reinforcement and Visualization in Sparsely-Matched Self-Organizing Maps
IEEE Transactions on Neural Networks
A self-organizing map (SOM) is a self-organized projection of high-dimensional data onto a typically 2-dimensional (2-D) feature map, wherein vector similarity is implicitly translated into topological closeness in the 2-D projection. However, when there are more neurons than input patterns, it can be challenging to interpret the results, due to diffuse cluster boundaries and limitations of current methods for displaying interneuron distances. In this brief, we introduce a new cluster…
A self-organizing map (SOM) is a self-organized projection of high-dimensional data onto a typically 2-dimensional (2-D) feature map, wherein vector similarity is implicitly translated into topological closeness in the 2-D projection. However, when there are more neurons than input patterns, it can be challenging to interpret the results, due to diffuse cluster boundaries and limitations of current methods for displaying interneuron distances. In this brief, we introduce a new cluster reinforcement (CR) phase for sparsely-matched SOMs. The CR phase amplifies within-cluster similarity in an unsupervised, data-driven manner. Discontinuities in the resulting map correspond to between-cluster distances and are stored in a boundary (B) matrix. We describe a new hierarchical visualization of cluster boundaries displayed directly on feature maps, which requires no further clustering beyond what was implicitly accomplished during self-organization in SOM training. We use a synthetic benchmark problem and previously published microbial community profile data to demonstrate the benefits of the proposed methods.
Find the Matlab code for implementing the above invention in the following url: http://www.mathworks.com/matlabcentral/fileexchange/35538-cluster-reinforcement-cr-phaseOther authorsSee publication -
Evolutionary mining for multivariate associations in large time-varying data sets: a healthcare network application
We introduce a new method for exploratory analysis of large data sets with time-varying features, where the aim is to automatically discover novel relationships between features (over some time period) that are predictive of any of a number of time-varying outcomes (over some other time period). Using a genetic algorithm, we co-evolve (i) a subset of predictive features, (ii) which attribute will be predicted (iii) the time period over which to assess the predictive features, and (iv) the time…
We introduce a new method for exploratory analysis of large data sets with time-varying features, where the aim is to automatically discover novel relationships between features (over some time period) that are predictive of any of a number of time-varying outcomes (over some other time period). Using a genetic algorithm, we co-evolve (i) a subset of predictive features, (ii) which attribute will be predicted (iii) the time period over which to assess the predictive features, and (iv) the time period over which to assess the predicted attribute. After validating the method on 15 synthetic test problems, we used the approach for exploratory analysis of a large healthcare network data set. We discovered a strong association, with 100% sensitivity, between hospital participation in multi-institutional quality improvement collaboratives during or before 2002, and changes in the risk-adjusted rates of mortality and morbidity observed after a 1-2 year lag. The results provide indirect evidence that these quality improvement collaboratives may have had the desired effect of improving health care practices at participating hospitals. The proposed approach is a potentially powerful and general tool for exploratory analysis of a wide range of time-series data sets.
Other authors -
The Discussion of Possibilities of Applications of Method of Linear and Nonlinear Wave Dynamics to Probability Determination in Wondering Problems
Proceedings of International conference on Security of Information Systems
Fun paper to read if you are into math and physics.
Honors & Awards
-
Faculty Choice Prize
Computer Science Department, University of Vermont
http://www.uvm.edu/~cems/cs/csfair/2013/fairover/home.php
Languages
-
English
Full professional proficiency
-
Russian
Full professional proficiency
-
Armenian
Native or bilingual proficiency
Recommendations received
-
LinkedIn User
5 people have recommended Narine
Join now to viewMore activity by Narine
-
I just learned about this product, but apparently students are already all over it. Record a lecture and Turbolearn will make flashcards, quizzes…
I just learned about this product, but apparently students are already all over it. Record a lecture and Turbolearn will make flashcards, quizzes…
Liked by Narine Hall, Ph.D.
-
So true. We want AI to take away the shitty labour we don't want to do. Not the stuff we love, but don't have enough time for. Not the stuff that…
So true. We want AI to take away the shitty labour we don't want to do. Not the stuff we love, but don't have enough time for. Not the stuff that…
Liked by Narine Hall, Ph.D.
-
Before I share beautiful pictures from yesterday, I’d like to first thank my managers who put this magical event together for our student. Their…
Before I share beautiful pictures from yesterday, I’d like to first thank my managers who put this magical event together for our student. Their…
Liked by Narine Hall, Ph.D.
-
Say my name? Say my name? Don't. I ditched the boring "Hi, I'm Anna Bohonek" intros and learned to set the stage with a story and prioritize the…
Say my name? Say my name? Don't. I ditched the boring "Hi, I'm Anna Bohonek" intros and learned to set the stage with a story and prioritize the…
Liked by Narine Hall, Ph.D.
-
LinkedIn is cancer, because everyone is encouraging you to be fake, pretentious and prioritize quantity over quality. Would prefer a topic with…
LinkedIn is cancer, because everyone is encouraging you to be fake, pretentious and prioritize quantity over quality. Would prefer a topic with…
Liked by Narine Hall, Ph.D.
-
May was an intense month filled with sports events that provided valuable lessons applicable to business. Here are some key takeaways: First, I…
May was an intense month filled with sports events that provided valuable lessons applicable to business. Here are some key takeaways: First, I…
Liked by Narine Hall, Ph.D.
-
I’m excited to announce I have been selected as sole finalist to become the 17th president of the University of North Texas! I’m grateful to…
I’m excited to announce I have been selected as sole finalist to become the 17th president of the University of North Texas! I’m grateful to…
Liked by Narine Hall, Ph.D.
Other similar profiles
Explore collaborative articles
We’re unlocking community knowledge in a new way. Experts add insights directly into each article, started with the help of AI.
Explore More