Private predictions on hidden Markov models

Abstract

Hidden Markov models (HMMs) are widely used in practice to make predictions. They are becoming increasingly popular models as part of prediction systems in finance, marketing, bio-informatics, speech recognition, signal processing, and so on. However, traditional HMMs do not allow people and model owners to generate predictions without disclosing their private information to each other. To address the increasing needs for privacy, this work identifies and studies the private prediction problem; it is demonstrated with the following scenario: Bob has a private HMM, while Alice has a private input; and she wants to use Bob's model to make a prediction based on her input. However, Alice does not want to disclose her private input to Bob, while Bob wants to prevent Alice from deriving information about his model. How can Alice and Bob perform HMMs-based predictions without violating their privacy? We propose privacy-preserving protocols to produce predictions on HMMs without greatly exposing Bob's and Alice's privacy. We then analyze our schemes in terms of accuracy, privacy, and performance. Since they are conflicting goals, due to privacy concerns, it is expected that accuracy or performance might degrade. However, our schemes make it possible for Bob and Alice to produce the same predictions efficiently while preserving their privacy.

Stock Selection through Hidden Markov Model: A Case of Pakistan Stock Exchange

“…Different authors used HMMs for predictions in various fields. Moreover, HMM does not disclose private data (Polat, Du, Renkes, & Oysal, 2010).…”

Detection and Classification of Skin Disease Using Modified Mobilenet Architecture

“…Khan et al (13) has worked on HAM10000 dataset using the Convolutional Neural Network (CNN) and has obtained an accuracy of 86.50% whereas Masni (14) et al has worked on a set of 2750 dermoscopy images using pre-trained architectures and has obtained different values of accuracy approximately 82% for all the models. Authors has used HAM10000 dataset (15)(16) using CNN model and has obtained values of accuracy as 85.80% and 77.00%. Salian et al (17) and More et al (18) has used a total of 10015 dermoscopy images using Mobilenet architecture and has obtained values of accuracy as 81.52% and 75.03% respectively.…”