Lightweight Deep Learning-Based DDoS Detection in SDN-VANET Using LSTM Autoencoder

@sajjadtaghvaeifr

Hello, I trust you're doing well. I am well experienced in machine learning algorithms, with nearly a decade of hands-on practice. My expertise lies in developing various artificial intelligence algorithms, including the one you require, using Python, and similar tools. I have worked with pytorch, and tensorflow to develop DL models, .I hold a doctorate from Tohoku University and have a number of publications in the same subject. My portfolio, which showcases my past work, is available for your review. Your project piqued my interest, and I would be delighted to be part of it. Let's connect to discuss in detail. Warm regards. please check my portfolio link: https://www.freelancer.com/u/sajjadtaghvaeifr

@saba106

I am an expert statistician, Research Writer, and data analyst with more than eight years of experience. I have full command of Excel analysis, SPSS, STATA, R LANGUAGE, AND PYTHON. I am an expert in creating time series prediction models, working with survey data, conducting marketing analysis, building estimators, and medical analysis. I am a perfect match for your project share other details of the work so I can start working on your project. Will complete task on time.

@Jennycreation

Hello Sir, I have already worked on similar types of website and software development projects and I am confident that I can successfully handle your project. I am a full-time freelancer with **17+ years of experience** in web development, software development, automation, and API integrations. My expertise includes: * ASP.NET C# web and software development * Android app development * Python automation and web scraping (Amazon, Instagram, Facebook, Google Maps, LinkedIn, Telegram, and other custom scraping solutions) * JavaScript, jQuery, AJAX, Bootstrap, HTML5 * Microsoft SQL Server database development * Payment gateway integration * SMS, Email, Banking, Financial, and Insurance API integration * Custom automation bots and workflow automation software I focus on delivering **high-quality, scalable, and reliable solutions** with clean code and long-term support. Please contact me in private chat so we can discuss your requirements in detail. I can also show you a **demo of similar projects** I have completed. Looking forward to working with you. Thanks & Regards

@FreeLancers360

Hi there, Your project "Lightweight Deep Learning-Based DDoS Detection in SDN-VANET Using LSTM Autoencoder" instantly caught my attention — it's exactly the kind of work we love to take on. What stood out to me is how clearly you’ve described what you want — and clarity like that is where we deliver our very best work. From your brief I can see this involves machine learning — all areas we handle in-house. We specialise in C Programming, Python, CUDA, Machine Learning (ML), which lines up directly with what you need. How we'd approach it: - Clarify the use-case, inputs and the exact output you expect - Build and integrate the model / automation pipeline - Evaluate accuracy and tune against real examples - Deploy with monitoring and a clear handover I also noted you don’t need explicitly documenting and obtaining appr — that’s clear, and I’ll keep the work clean and strictly on-brief. Delivering at the scale of 20 the is no problem for us — we're set up for volume without dropping quality. Expect smooth progress updates throughout, full respect for your specifications, and zero surprises along the way. Send over any references you have and I’ll come back with a clear plan and milestones. Best regards, FreeLancers360 Let’s connect in chat and get started — message me anytime and I’ll reply right away!

@InjectAI

I can implement this research workflow using the actual CICDDoS2019 dataset with strict reproducibility and data leakage prevention. My background includes Python, PyTorch, scikit-learn, pandas, machine learning, deep learning, and model evaluation, so I can handle both the LSTM-AE methodology and the comparative experiments. I will first audit and document the supplied dataset files, samples, labels, attack categories, selected features, missing/infinite values, class distribution and train/validation/test split. The LSTM Autoencoder will be trained primarily on normal traffic, with sequences created using a documented sliding-window strategy and the detection threshold selected only from validation data. Deliverables: • Complete CICDDoS2019 preprocessing pipeline • LSTM Autoencoder implementation • Random Forest, SVM, KNN, ANN and CNN-LSTM baselines • Fair and reproducible experimental splits • Accuracy, Precision, Recall, F1, ROC-AUC, FPR and FNR • Confusion matrices, ROC and PR curves • Reconstruction-error and threshold analysis • Feature-selection and sequence-length experiments • Threshold sensitivity analysis • Training/inference time and parameter analysis • Numerical result tables and publication-ready figures • Complete source code, requirements and configurations • Saved models, results and reproducibility documentation

@Senavirathna1014

Hi, I can implement a reproducible CICDDoS2019 study covering your LSTM Autoencoder and all five baselines. I will report measured results honestly, without promising that LSTM-AE will outperform other models. I will audit the actual files and document sample counts, benign/attack distributions, categories, features, labels, and missing/infinite-value handling. No synthetic data, oversampling, or unapproved methodological changes will be introduced. To prevent leakage, I will partition data before constructing windows, preserve temporal order, prevent sequences crossing split boundaries, and fit preprocessing only on training data. LSTM-AE training will use normal traffic; tuning, sequence-length selection, and threshold selection will use validation data only. RF, SVM, KNN, ANN, and CNN-LSTM will receive fair tuning on consistent partitions where appropriate, with differences documented. I will deliver the requested metrics, efficiency measurements, tables, curves, confusion matrices, and reconstruction-error analysis. Subject to compute capacity, I will compare full numerical features against the specified 13 features, sequence lengths 5/10/20, and threshold sensitivity. Delivery includes source code, environment requirements, scripts, configurations, seeds, saved models, results, figures, and execution instructions.

@ronakkataria1

One line in your brief decides whether the results hold up: training the autoencoder primarily on normal traffic. With CIC-DDoS2019 that is harder than it sounds. The benign share is small, and several attack files carry almost no normal flows at all. So the normal-only training set has to be pooled across files and kept strictly apart from the benign samples used for validation and test. Otherwise the threshold is tuned on traffic the model has already seen and the detection rate comes out optimistic. Two things change the numbers more than the architecture does. The dataset carries socket columns, source and destination IP, port and timestamp, and leaving them in lets the model recognise attacks by address rather than behaviour. That reads as excellent and generalises to nothing. And the sliding window has to respect flow boundaries, or sequences straddle two unrelated flows. For the threshold: reconstruction error on validation traffic that is normal only, percentile fixed before any test data is touched, with the value and the rule written down. At this budget I would scope phase one as the LSTM autoencoder end to end with the documentation you listed, and add the Random Forest and SVM baselines as a second step. Which CIC-DDoS2019 files do you have, the full set or the day folders? Ronak

@bilal6901

I understand the goal is to build a reproducible LSTM Autoencoder-based DDoS detection pipeline using the actual CIC-DDoS2019 dataset, with fair baseline comparisons and proper leakage prevention. I can work through the project systematically: • Dataset inspection and detailed preprocessing • LSTM Autoencoder anomaly-detection pipeline • Random Forest, SVM, KNN, ANN and CNN-LSTM baselines • Accuracy, precision, recall, F1, ROC-AUC, FPR and FNR • Reconstruction-error and threshold analysis • Feature-selection and sequence-length experiments • Reproducible Python code, results, figures and tables • Clear documentation of splits, preprocessing and leakage prevention I will keep the experiments reproducible and document the major methodological decisions clearly. I can start immediately.

@liamc081

Hi, Liam here from Dallas. I understand you need a reproducible CIC-DDoS2019 study using the actual dataset and an LSTM Autoencoder for DDoS anomaly detection. You also need fair baseline models, detailed metrics, ablation tests, and complete Python code. The key priorities are preventing data leakage and keeping training, validation, and testing properly separated. Dataset details, feature selection, threshold selection, and model comparisons must also be clearly documented. I’ll build the preprocessing, LSTM-AE, baseline models, evaluation pipeline, figures, and tables with fixed seeds. I’ll also provide the source code, requirements, saved models, results, and documentation for reproducibility. I have experience with Python machine learning, deep learning, LSTM models, and anomaly detection. I’m also familiar with reproducible experiments, preprocessing, validation, and leakage-safe model evaluation. A few questions to better understand the scope: Q1. Can you provide the CIC-DDoS2019 files and 13-feature specification? Q2. Do you have a required train/validation/test split? Q3. Do you prefer TensorFlow or PyTorch? Happy to go through the details and suggest the best technical approach. Looking forward to hearing from you.

@Youssefkamel123

Hello, I’m a Machine Learning developer experienced in Python, data preprocessing, model training, evaluation, and reproducible ML workflows. I can implement your CICDDoS2019 DDoS detection experiment using the actual dataset files and fully document the dataset, features, labels, class distribution, preprocessing, and train/validation/test split. The proposed LSTM Autoencoder will be trained primarily on normal traffic using normalized features and sliding-window sequences. Reconstruction error and the final detection threshold will be determined using validation data only. I will strictly prevent test-set leakage during preprocessing, feature selection, scaling, tuning, sequence construction, and threshold selection. I will also implement and fairly compare: • Random Forest • SVM • KNN • ANN • CNN-LSTM • LSTM Autoencoder Evaluation will include Accuracy, Precision, Recall, F1, ROC-AUC, FPR, FNR, confusion matrices, training/inference time, and model complexity where applicable. For LSTM-AE, I will provide reconstruction-error plots, threshold analysis, ROC and PR curves, confusion matrix, and threshold sensitivity. I can also evaluate the 13-feature set and sequence lengths 5, 10, and 20 where computationally feasible. Deliverables: complete Python code, preprocessing/training/evaluation scripts, requirements, saved models/results, figures, tables, configurations, and reproducibility details.

@Tradeware

Hi, I can implement this study using the actual CIC-DDoS2019 dataset with a reproducible, leakage-free experimental pipeline. I’ll build the LSTM Autoencoder anomaly detector plus Random Forest, SVM, KNN, ANN and CNN-LSTM baselines using consistent splits. I’ll document preprocessing, class distribution, feature selection, sliding windows, validation-based threshold selection and all hyperparameters. Results will include Accuracy, Precision, Recall, F1, ROC-AUC, FPR/FNR, confusion matrices, timing/efficiency metrics, required plots and ablation experiments. You’ll receive complete Python code, requirements, seeds/configuration, trained models, figures, tables and reproducibility documentation. Keywords: CICDDoS2019, Python, LSTM Autoencoder, Anomaly Detection, CNN-LSTM, Machine Learning, Deep Learning, Data Leakage Prevention, Reproducibility

@Shiksha52

Hi, This project strongly aligns with my research in intrusion detection and network security. I have worked with UNSW-NB15, NSL-KDD and CICIDS2017 on feature processing, Random Forest/CNN-based detection, threshold calibration and rigorous security-model evaluation. I would build this as a controlled anomaly-detection experiment rather than optimizing for a high test score. I would first establish the dataset pipeline and freeze the test set, fit preprocessing only on training data, and use a separate validation stage to select the LSTM-AE reconstruction threshold. This keeps the final results genuinely unbiased. For the comparative experiments, I would maintain a consistent split and preprocessing strategy and structure the code so that feature reduction, sequence length and threshold sensitivity can be evaluated independently. I would also save seeds, configurations and intermediate results so the reported metrics, tables and figures can be reproduced directly. My previous IDS research includes adaptive thresholding, class-imbalance handling and security-oriented model evaluation; one of my frameworks achieved 99.71% precision and 99.48% recall. I can deliver the complete implementation, experiments, results and research-ready documentation. Before starting, I would confirm the exact 13-feature definition and CICDDoS2019 file structure to avoid making assumptions about the intended experiment. Best regards, Shiksha

@ansara05

Hi, I have hands-on experience with machine learning, deep learning, anomaly detection, network security datasets, LSTM Autoencoders, and comparative model evaluation. I can work directly with the actual CICDDoS2019 dataset and keep the entire experiment reproducible and leakage-free. I’ll build the LSTM-AE pipeline with proper preprocessing, sequence generation, normal-traffic training, validation-based threshold selection, and final evaluation on an untouched test set. I’ll also implement RF, SVM, KNN, ANN and CNN-LSTM baselines using a fair experimental setup. I can handle the 13-feature comparison, sequence-length experiments, threshold sensitivity analysis, required metrics, confusion matrices, ROC/PR curves, reconstruction-error plots, tables and figures. You’ll receive clean Python code, saved models/results, requirements, and clear documentation of the methodology and experimental decisions.

@vijayd091

Hello, I can implement and evaluate the complete DDoS anomaly-detection pipeline using the actual CICDDoS2019 / CIC-DDoS2019 dataset, with a strong focus on reproducibility, data-leakage prevention, and fair model comparison. What I will deliver Complete CICDDoS2019 dataset preprocessing pipeline Clear documentation of dataset files, samples, benign/DDoS distribution, attack categories, features, labels, and missing/infinite values Reproducible train/validation/test split Proper normalization/scaling fitted only on training data Sliding-window temporal sequence construction LSTM Autoencoder trained primarily on normal traffic Reconstruction-error calculation and validation-based threshold selection Final DDoS/normal classification and evaluation Comparative Models I will implement and evaluate: Random Forest SVM KNN ANN CNN-LSTM LSTM Autoencoder