Debasmita Morgan Email & Phone Number

Responsibilities:• Provide statistical support for research and product development.• Develop experimental designs for consumer product testing, and other applications.• Analyze databases to identify emerging trends and potential problems.• Perform Natural Language Processing to mine text data.• Build RShiny apps to facilitate automation of product/appliance analyses.

Stamford, Connecticut, United States

Storrs Center, Connecticut, United States

Responsibilities:• Provided instructional support in statistics for students across all programs at UConn. Illustrated key points using quantitative reasoning.• Attempted to expand students’ theoretical understanding of statistics and mathematical foundation, and tried to develop their abilities in exposition and problem solving.• Engaged with the students on a one-one level to foster a spirit of inquiry and intellectual curiosity. Achievements:Received excellence in teaching certificates in 2018 and 2019 from University of Connecticut.

Kolkata, India

Project title: Robust Speaker IdentificationBackground: Automatic speaker identification/recognition (ASI/ASR) is the generic term applied to the automatic process of inferring the identity of a person from an utterance made by him, on the basis of speaker-specific information embedded in the corresponding speech signal. This technique can be used to verify the identity claimed by people accessing systems, that is, it enables access control of various services by voice. Real-life activities where it is immediately applicable and useful include voice dialing, banking over a telephone network, telephone shopping, database access services, information and reservation services, voice mail, security control for confidential information, and remote access to computers. Another important application of speaker recognition technology is its use for forensic purposes.My contribution:• Analyzed the closed set, text-independent variant of the speaker recognition problem.• Applied principal component transformation to the feature vectors of the corresponding speaker to make them uncorrelated.• Proposed a novel solution to the speaker identification problem through minimization of statistical divergences between the probability distribution of feature vectors from the test utterance and the probability distribution of the feature vectors corresponding to the speaker classes.Achievements:• Significant improvement in classification accuracy was observed under the proposed approach on the benchmark speech corpus.• Further, the ubiquitous principal component transformation, by itself and in conjunction with the principle of classifier combination, improved the performance by 15%.

Ms, Statistics

Master's Project - Comparison of classification techniques for drug response micro array data • Analyzed drug responses micro array gene.

Ms, Statistics

Bs, Statistics