Comparison of CNN, LSTM and Their Hybrid Models in Detecting Coronavirus Using Genome Sequences
- سال انتشار: 1403
- محل انتشار: دومین کنفرانس بین المللی هوش مصنوعی و فناوری های آینده نگر
- کد COI اختصاصی: ICAIFT02_026
- زبان مقاله: انگلیسی
- تعداد مشاهده: 42
نویسندگان
Dept. of Electrical and Computer Engineering, Tarbiat Modares University, Tehran, Iran
Dept. of Biomedical Engineering, Tehran Medical Sciences Islamic Azad University, Tehran, Iran
Dept. of Biomedical Engineering, Tehran Medical Sciences Islamic Azad University, Tehran, Iran
چکیده
Precise and expeditious identification of harmful viruses, such as the coronavirus, is essential for managing the transmission of diseases and averting worldwide pandemics. Several techniques for identifying the coronavirus, including chest radiography, RT-qPCR testing, and antibody tests, are available. Nevertheless, these procedures, with the exception of their time-consuming nature, are incapable of detecting the virus from the moment it enters the patient’s body. Moreover, none of these techniques can accurately detect and categorize the many strains of coronaviruses found within the patient's body. As a result, novel diagnostic techniques utilizing viral genome sequences and neural networks have been developed through the application of deep learning. The method employed for identifying and classifying viruses involves the utilization of convolutional networks. In addition, recurrent neural networks are used for classifying DNA strands and genome sequences due to their sequential structure. This research begins by analyzing two models for the classification of coronaviruses based on genomic data: a Convolutional Neural Network (CNN) and Long Short-Term Memory (LSTM). The CNN and LSTM classifier models attained an accuracy rate of ۸۴.۶۲% and ۹۰.۳۸%, respectively, in the identification and classification of viruses. Research showed that the integration of these two models may enhance the performance of both networks, leading to higher accuracy rates. The LSTM neural network in these hybrid models utilizes features collected by the CNN for classification purposes. A novel hybrid model, known as CNN۱D-LSTM, is presented to enhance accuracy by using the strengths of both CNN and LSTM models. This model achieved an impressive accuracy of ۹۶.۱۵%. Ultimately, in order to analyze two-dimensional data, a model called CNN۲D-LSTM, which combines two-dimensional CNN and LSTM, is introduced. This model achieved a maximum accuracy of ۹۸.۰۸% in identifying different types of coronaviruses. The findings demonstrate that the proposed models can be effective instruments for virus identification and classification and can greatly assist in addressing worldwide epidemics.کلیدواژه ها
genome sequence, neural networks, deep learning, CNN, LSTM, virus identification and classificationمقالات مرتبط جدید
- تحلیل چالشها و راهکارهای تقویت ارتباط دانشگاه و صنعت: با تمرکز بر حلقههای مفقوده
- بازخوانی نقش دانشگاه و صنعت در توسعه ملی: از موانع تا راهکارها
- نشانگر تشخیصی جدید در ژن C-myc به عنوان کیت غیر تهاجمی تشخیص سرطان دهان
- برنامه ریزی منابع تجدید پذیر با درنظر گرفتن برنامه ریزی توسعه انتقال و تولید منابع توان راکتیو
- برنامه ریزی همزمان توسعه انتقال و منابع تولید توان راکتیو با استفاده از یک الگوریتم تکاملی بهبود یافته
اطلاعات بیشتر در مورد COI
COI مخفف عبارت CIVILICA Object Identifier به معنی شناسه سیویلیکا برای اسناد است. COI کدی است که مطابق محل انتشار، به مقالات کنفرانسها و ژورنالهای داخل کشور به هنگام نمایه سازی بر روی پایگاه استنادی سیویلیکا اختصاص می یابد.
کد COI به مفهوم کد ملی اسناد نمایه شده در سیویلیکا است و کدی یکتا و ثابت است و به همین دلیل همواره قابلیت استناد و پیگیری دارد.