Browsing by Author "Chintawar, S."

Filter results by typing the first few letters
Now showing 1 - 3 of 3
  • No Thumbnail Available
    Item
    Multi-Level Statistical Model for Forecasting Solar Radiation
    (Institute of Electrical and Electronics Engineers Inc., 2022) Nayak, P.; Dash, A.; Chintawar, S.; Anand Kumar, M.
    As a substitute for conventional energy sources, Solar energy is quickly becoming a popular source of renewable energy. Various entities ranging from small households and businesses to large firms and MNCs are currently making plans on investing resources in the generation of solar energy. Thus, accurate prediction of solar radiation has become a necessity in the present scenario. Due to limitations like the unavailability of proper measuring equipment and a small number of meteorological departments, accurate prediction of solar radiation is not possible in many places around the world. This paper focuses on forecasting solar radiation using machine learning techniques. Solar radiation depends upon various natural factors, which are easier to measure, and these factors can help forecast solar radiation. This paper explores the available data to identify the various factors which affect solar radiation. Based on these factors, the paper investigates the performance of different standard regression models based on solar radiation prediction. Next, multi-level statistical models are proposed, which stack multiple standard models into layers, and the R2 scores of these custom models is compared with the R2 scores of the standard models. © 2022 IEEE.
  • No Thumbnail Available
    Item
    OntoPred: An Efficient Attention-Based Approach for Protein Function Prediction Using Skip-Gram Features
    (Springer, 2023) Chintawar, S.; Kulkarni, R.; Patil, N.
    Proteins play an essential role in performing many cellular functions in organisms and are responsible for various biochemical activities. The main objective of this task of protein function prediction is to annotate protein sequences with their correct functions, which are represented by Gene Ontology (GO) terms. Recently, the number of new proteins released has been increasing. As the experimental approach of annotating these proteins is very time-consuming, the need for faster annotation techniques has arisen. Approaches using deep learning and machine learning have been shown to be beneficial in this regard. In this research, we propose a novel approach, OntoPred, for the task of function prediction which makes use of the standalone protein sequences and annotates them with their corresponding functions (GO terms). The core idea is to use an attention mechanism to identify which parts of a sequence influence the presence of a function. The model uses a combination of n-grams and skip-gram features extracted from the sequences. The proposed model was evaluated on multiple datasets including the CAFA3 evaluation benchmark. The maximal F1 scores obtained on molecular function (MF), biological process (BP), and cellular component (CC) aspect on the CAFA3 evaluation benchmark are 0.494, 0.480, and 0.637 respectively. © 2023, The Author(s), under exclusive licence to Springer Nature Singapore Pte Ltd.
  • No Thumbnail Available
    Item
    Transformer and Knowledge Based Siamese Models for Medical Document Retrieval
    (Institute of Electrical and Electronics Engineers Inc., 2023) Dash, A.; Merchant, A.M.; Chintawar, S.; Kamath S․, S.
    Vocabulary mismatch is a significant issue when it comes to query-based document retrieval in the medical field. Since the documents are typically authored by professionals, they may contain many specialized terms that are not widely understood or used. Traditional information retrieval (IR) models like vector space and best match-based models fail in this regard. Neural Learning to Rank (NLtR) and transformer models have attracted significant research attention in the field of IR. Recent works in the medical field utilize medical knowledge bases (KB) that map words to concepts and aid in connecting several words to the same concept. In this paper, we present various Siamese-structured transformer and knowledge-based retrieval models designed to address the retrieval issues in the medical domain. The experimental evaluation highlighted the superior performance of the proposed retrieval model, and the best one, based on the UMLSBert ENG transformer, achieved best-in-class performance with respect to all evaluation metrics. © 2023 IEEE.