مطالب مرتبط با کلیدواژه

Random Forest


۱.

معرفی یک مدل جدید ترکیبی الگوریتم مبنا به منظور پیش بینی حساسیت زمین لغزش های سطحی اطراف شهر بیجار(مقاله علمی وزارت علوم)

نویسنده:
منبع: جغرافیا و توسعه بهار ۱۳۹۶ شماره ۴۶

کلیدواژه‌ها: الگوریتم بیجار زمین لغزش سطحی Random Forest Random Subspace

حوزه های تخصصی:
تعداد بازدید : ۱۳۹۱ تعداد دانلود : ۶۴۵
افزایش صحت و اعتماد و در نتیجه کاهش عدم قطعیت نقشه های پیش بینی مکانی مخاطرات زمینی از جمله زمین لغزش ها یکی از چالش های پیش رو در این گونه مطالعات می باشد. هدف این پژوهش ارائه یک مدل ترکیبی جدید داده کاوی الگوریتم- مبنا به نام Random Subspace-Random Forest (RS-RF)،برای افزایش میزان صحت پیش بینی مناطق حساس به وقوع زمین لغزش های سطحی اطراف شهر بیجار می باشد. در ابتدا، نوزده عامل مؤثر بر وقوع زمین لغزش های سطحی منطقه ی مورد مطالعه شامل درجه شیب، جهت شیب، ارتفاع از سطح دریا، انحنای معمولی شیب(Curvature)، تقعر و تحدب شیب(Profile curvature)، همگرایی و واگرایی شیب (Plan curcvature)، شدت تابش خورشید (Solar radiation)، شاخص قدرت جریان، شاخص نمناکی توپوگرافی، شاخص طول و زاویه شیب، کاربری ارضی، شاخص پوشش گیاهی، لیتولوژی، فاصله از گسل، تراکم گسل، بارندگی، فاصله از آبراهه، تراکم آبراهه و فاصله از شبکه جاده شناسایی شدند. سپس، بر اساس شاخص Information Gain Ratioدوازده عامل مؤثر از بین آن ها انتخاب و جهت مدل سازی به کار گرفته شدند. اهمیّت نسبی هر کدام از عوامل در مدل Random Forest و مدل ترکیبیRS-RFبررسی شدند.معیارهای Kappa، Precision، Recall، F-Measure، AUROCبرای ارزیابی مدل ها هم برای داده های تعلیمی و هم برای داده های صحت سنجی استفاده شدند. نقشه های پیش بینی مکانی وقوع زمین لغزش های سطحی با این دو مدل نیز به دست آمدند. نتایج نشان داد که در مدل RF جهت شیب و در مدل ترکیبی RS-RFدرجه شیب مهم ترین فاکتورهای مؤثر بر وقوع زمین لغزش های منطقه ی مورد مطالعه شناخته شدند. نتایج ارزیابی مدل توسط معیارهای معرفی شده بیانگر تأیید این مدل ها برای داده های تعلیمی و داده های صحت سنجی بودند. نتایج ارزیابی صحت نقشه پهنه بندی به دست آمده نشان داد که درصد مساحت زیر منحنیROC(AUROC) برای داده های تعلیمی در مدل RF و مدل ترکیبی RS-RFارائه شده به ترتیب 729/0 و 784/0 وبرای داده های صحت سنجی به ترتیب 717/0 و 771/0 به دست آمدند. بطور کلی، نتایج نشان داد که تکنیک Random Subspaceمنجر به افزایش صحت پیش بینی مکانی حساسیت زمین لغزش های سطحی منطقه ی مورد مطالعه شده است. دستیابی به یک نقشه ی پیش بینی مکانی زمین لغزش های سطحی با صحت بالاتر، کمک شایانی در توسعه ی معقول تر تأسیسات، اراضی شهری و روستایی، طرح های آمایش سرزمین، طرح های آبخیزداری و همچنین جلوگیری از هدر رفت خاک و فرسایش توده ای و انتقال رسوبات به پایین دست خواهد شد.
۲.

Machine Learning Algorithms Performance Evaluation for Intrusion Detection(مقاله علمی وزارت علوم)

نویسنده:
منبع: Journal of Information Technology Management , Volume ۱۳, Issue ۱, ۲۰۲۱ 42 - 61

کلیدواژه‌ها: Intrusion Detection System Naïve Bayes Random Forest Support vector machine

حوزه های تخصصی:
تعداد بازدید : ۳۹۰ تعداد دانلود : ۱۳۱
The steadily growing dependency over network environment introduces risk over information flow. The continuous use of various applications makes it necessary to sustain a level of security to establish safe and secure communication amongst the organizations and other networks that is under the threat of intrusions. The detection of Intrusion is the major research problem faced in the area of information security, the objective is to scrutinize threats or intrusions to secure information in the network Intrusion detection system (IDS) is one of the key to conquer against unfamiliar intrusions where intruders continuously modify their pattern and methodologies. In this paper authors introduces Intrusion detection system (IDS) framework that is deployed over KDD Cup99 dataset by using machine learning algorithms as Support Vector Machine (SVM), Naïve Bayes and Random Forest for the purpose of improving the precision, accuracy and recall value to compute the best suited algorithm.
۳.

Investigating the Role of Code Smells in Preventive Maintenance(مقاله علمی وزارت علوم)

نویسنده:
منبع: Journal of Information Technology Management , Volume ۱۱, Issue ۴, ۲۰۱۸ 41 - 63

کلیدواژه‌ها: Preventive maintenance Code smells Machine Learning Random Forest

حوزه های تخصصی:
تعداد بازدید : ۲۴۲ تعداد دانلود : ۹۲
The quest for improving the software quality has given rise to various studies which focus on the enhancement of the quality of software through various processes. Code smells, which are indicators of the software quality have not been put to an extensive study for as to determine their role in the prediction of defects in the software. This study aims to investigate the role of code smells in prediction of non-faulty classes. We examine the Eclipse software with four versions (3.2, 3.3, 3.6, and 3.7) for metrics and smells. Further, different code smells, derived subjectively through iPlasma, are taken into conjugation and three efficient, but subjective models are developed to detect code smells on each of Random Forest, J48 and SVM machine learning algorithms. This model is then used to detect the absence of defects in the four Eclipse versions. The effect of balanced and unbalanced datasets is also examined for these four versions. The results suggest that the code smells can be a valuable feature in discriminating absence of defects in a software.
۴.

Detection of Wormhole Attack in Vehicular Ad-hoc Network over Real Map using Machine Learning Approach with Preventive Scheme(مقاله علمی وزارت علوم)

نویسنده:
منبع: Journal of Information Technology Management , Volume ۱۴, Security and Resource Management challenges for Internet of Things, ۲۰۲۲ 159 - 179

کلیدواژه‌ها: VANET AODV Broadcast Unicast k-NN Random Forest SUMO-0.32.0 NS-3.24.1 Packet leash Cryptography

حوزه های تخصصی:
تعداد بازدید : ۳۳۷ تعداد دانلود : ۱۲۵
VANET (Vehicular Ad-hoc Network) is a developing technology, which is a combination of cellular technology, ad-hoc network & wireless LAN to improve the safety of vehicle as well as driver. VANET communication can be of two types, first one is broadcast and second one is unicast. Either communication may be broadcast or unicast both are sensitive to different types ofassaults, for example message forgery, (DOS) denial of service, Sybil assault, Greyhole, Blackhole & Wormhole assault. In this paper machine learning method is used to detect the wormhole assault in VANET’s multi-hop communication. We have created a scenario of VANET by using AODV routing protocol on NS-3.24.1 simulator, which utilizes the overall mobility traces generated by the simulator SUMO-0.32.0 to model the wormhole assault. The simulation is performed by using NS-3.24.1 simulator, and the statistics created by flow monitor are collected. The collected data is pre-processed and the k-NN & Random Forest algorithms are applied on this data, to make the model such type so that it can memorize the wormhole attack. The novelty of this research work is that with the help of proposed detection & prevention technique, vehicular ad-hoc network can be made free from wormhole assault by using ML approach. The performance of proposed machine learning models is compared with existing work. In this way it is clear that our proposed approach by using ML is powerful tool by which the wormhole assaults can be detected in VANETs. A scheme based on packet lease and cryptographic techniques is used to prevent the wormhole attack in VANET
۵.

Evaluation of flood potential of Ardabil plain using fuzzy models and satellite images

نویسنده:
منبع: جغرافیا و روابط انسانی دوره ۶ پاییز ۱۴۰۲ شماره ۲۲ 369 - 389

کلیدواژه‌ها: Ardabil Plain Flood Logistic regression Random Forest

حوزه های تخصصی:
تعداد بازدید : ۷۱ تعداد دانلود : ۶۷
Ardebil plain is one of the flood points that requires the understanding of the flood potential. In this study, the flooding potential of Ardebil plain was performed using environmental parameters, observations of flood points and lack of floods and prediction algorithms were made including random forest and logistics regression. Independent parameters include DEM, Slope, Aspect, Distance from waterway, distance from dam, runoff accumulation, land use, landforms and indexes Topographic Position Index (TPI), Modified Catchment Area (MCA), Terrain Ruggedness Index (TRI), Topographic Wetness Index (TWI) and Stream Power Index (SPI) Indices. The Roc-AUC assessment results showed that the RF and LR model were validated by 0.99 and 0.98, and it shows that random forest models and logistics regression have the ability to predict and prepare a flood sensitivity map in Ardebil plain. The output of parameters effective in flooding showed that the marginal areas located around the central plain of Ardabil have less flood-flooding potential than the central areas. The results also showed that by moving from the southwest of the plain to its northeast, the grade of floods increased. This increase in flooding potential around the main drainage of the plain is greater than elsewhere.
۶.

Comparative Study on Different Machine Learning Algorithms for Neonatal Diabetes Detection(مقاله علمی وزارت علوم)

نویسنده:
منبع: Journal of Information Technology Management , Volume ۱۶, Issue ۱, ۲۰۲۴ 5 - 26

کلیدواژه‌ها: Voting Classifiers Meta-Classification Technique Diabetes Risk Prediction Biomedical Clinical Risk Factors Random Forest Logistic regression Gradient Boosting Support Vector Machines

حوزه های تخصصی:
تعداد بازدید : ۸۷ تعداد دانلود : ۶۳
This paper gives a performance analysis of multiple vote classifiers based on meta-classification methods for estimating the risk of diabetes. The study's dataset includes a number of biological and clinical risk variables that can result in the development of diabetes. In the analysis, classifiers like Random Forest, Logistic Regression, Gradient Boosting, Support Vector Machines, and Artificial Neural Networks were used. In the study, each classifier was trained and evaluated separately, and the outcomes were compared to those attained using meta-classification methods. Some of the meta-classifiers used in the analysis included Majority Voting, Weighted Majority Voting, and Stacking. The effectiveness of each classifier was evaluated using a number of measures, including accuracy, precision, recall, F1-score, and Area under the Curve (AUC). The results show that meta-classification techniques often outperform solo classifiers in terms of prediction precision. Random Forest and Gradient Boosting, two different classifiers, had the highest accuracy, while Logistic Regression performed the worst. The best performing meta-classifier was stacking, which achieved an accuracy of 84.25%. Weighted Majority Voting came in second (83.86%) and Majority Voting came in third (82.95%).