Conference Papers

Permanent URI for this collectionhttps://idr.nitk.ac.in/handle/123456789/28506

Browse

Author: search.filters.author.Jena, D.Subject: search.filters.subject.K-nearest neighbor

Search Results

Now showing 1 - 2 of 2
  • No Thumbnail Available
    Item
    Comparison of two data cleaning methods as applied to volatile time-series
    (Institute of Electrical and Electronics Engineers Inc., 2019) Ranjan, K.G.; Prusty, B.R.; Jena, D.
    Out-of-sample forecasting of historically observed time-series inevitably necessitates the application of a suitable data cleaning method to assist improved accuracy of the obtained results. The existing data cleaning methods though work amply with nonvolatile time series; fail when applied to a volatile time-series. In this paper, the suitability of the k-nearest neighbor approach and sliding window prediction approach is tested on a set of nonvolatile and volatile time-series. The performance comparison is carried out considering the historical record of furniture sales data, PV generation, load power, and ambient temperature data of different time-steps collected from various places in the USA. Further, the effect of parameters allied with both the methods on the preprocessing result is also analyzed. Finally, possible reforms are suggested for the appropriate preprocessing of volatile time-series. © 2019 IEEE.
  • No Thumbnail Available
    Item
    Comparative analysis of different machine learning techniques for condition monitoring of capacitors in a SEPIC converter
    (Institute of Electrical and Electronics Engineers Inc., 2022) RAJENDRAN, S.; Jena, D.; Diaz-D, M.; Devi, V.S.K.
    An efficient condition monitoring technique is essential for power converters to avoid unscheduled maintenance. In this work, the condition monitoring of capacitors in a single-ended primary inductance converter (SEPIC) is proposed based on the following machine learning classifiers: K nearest neighbor, support vector machine, back propagation neural network, Naive Bayes, and deep neural network. The feature of the machine learning algorithms is extracted by three node voltages such as voltage across $C$ 1,C2, and load. These features are utilized for training the algorithms. Moreover, the effectiveness of the different classifies are evaluated by considering the accuracy and area under the curve. Further, each algorithm is trained with a different percentage of a dataset. Finally, a comparative study has been made between the algorithms, and the results exhibit that the deep neural network performs better classification than other algorithms. © 2022 IEEE.