This study presents an advanced system for detecting skin cancer using Convolutional Neural Networks (CNNs), enhanced by Batch Normalization to improve model stability during training. CNNs, widely recognized for their effectiveness in image analysis, form the foundation of this system, which is designed to address the global challenge of skin cancer detection. The model's capacity to manage a variety of datasets, providing enhanced adaptability, is one of its primary characteristics. It tackles the common issue of imbalanced skin cancer data by employing techniques such as SMOTE, undersampling, and oversampling, resulting in increased accuracy and sensitivity, particularly for less common cases. Comparative experiments demonstrate that this model surpasses previous benchmarks in identifying skin disorders. The integration of Group Normalization further boosts stability, and the combined methods for addressing data imbalances enhance the model's ability to generalize across varied data. This makes the system a highly valuable tool for healthcare professionals. Experimental evaluation on the HAM10000 dataset achieved a test accuracy of 96.4%, a training accuracy of 99.74%, and a validation accuracy of 96.35%, with a minimal loss of 0.0079. The adaptive data balancing strategy further enhanced classification, improving F1-scores by 12–15% for rare classes such as melanoma and dermatofibroma, while preserving 98.2% accuracy for majority classes. The study underscores the potential of modern deep learning techniques to transform the interpretation of medical images, setting a new standard to combating skin diseases in healthcare.

Lung cancer is one of the main cause of cancer related death globally so the early detection techniques are really essential for reducing the deaths globally every year. In order to diagnose lung cancer, our research proposes a hybrid machine learning technique which uses XGBoost for tabular data analysis and Convolutional Neural Networks (CNN) for histopathological image classification. The suggested approach makes use of two datasets: a histopathological image dataset with 25,000 images of lung cancer, and a lung cancer dataset with 15 clinical and demographic variables. While the CNN model showed 91.06% training and validation accuracy for image classification whereas the XGBoost model achieved 98.39% accuracy for tabular data set. we used Streamlit to create a web based tool for this research project for real time prediction. The hybrid approach is great for early lung cancer screening programs and early lung cancer detection and support systems.

It has been called to attention that road traffic crashes (RTCs) have a major social and economic impact, but risk of accident may be analyzed at an individual level of individual isolated occurrence. Eventually we realized that there is a significant trend behind this perception: Risk is not uniform. At certain places, it appears that the severity of accidents is (constantly) larger, even when a single crash appears to be identical. Then the observation at the latter resulted in the need to have a regional accident risk assessment score that is not pegged on a single event but a more broad environmental and demographic features. This research is aimed at the development of predictive risk models of accidents at the postal-code-level by considering concurrently the historic events of accidents, population compositions and road network characteristics. We do not position this as a pure classification task, but rather estimate the severity of the accident using ensemble machine learning models, including Random Forest, XGBoost, LightGBM, and CatBoost, and take the sum of the predictions to form a numerical and accident risk score in each postal code This aggregation-based idea has been shown to be more stable in locations with uneven accident frequency, which should be useful in addressing a region with uneven accident frequency. Regression-based metrics of model performance are used to compare model performance meaningfully. The reason why the system is not being turned into a black box is to explain the importance of features and how factors are comprehended through explainable machine learning methods.density of population and road network affect risk in regions like population density and road structure. Interestingly enough, features related to infrastructure had a tendency to dominate the simple count of accidents of purely historical nature. Altogether, this paper shows that the practical and interpretable method of accident risk analysis can help to assess the safety of the region and make the informed decisionmaking process, being based on the conditions of real-world data. Stable regional risk patterns have been shown in experiments on over 49,000 postal codes, even though there is only slight prediction accuracy on an accident level.

Early Prediction of Heart disease is a major concern now a days because of its mortality rate in the global society, which requires effective ways of detection. The increased availability of clinical and lifestyle information offers the possibility to implement the machine learning methods to help in the adequate diagnosis of a patient. We put forward a Hybrid Random Forest-Logistic Regression (HRFLM) algorithm that uses Random Forest for selection and Logistic regression for classification. A comparative analysis of different models along with the proposed method for heart disease prediction are monitored with respect to the feature-rich dataset of demographics, clinical, and behavioral factors. The Support Vector Machine, Decision Tree, Random Forest, and HRFLM are the tested ones. Upon data processing and cleaning, the models are evaluated based on different performance metrics. The experiments prove that the HRFLM Model works more effectively than regular machine learning models. The hybrid method had a high precision and a high reduction in false negatives of the tool, which proved to be an effective tool of clinical decision support in the early detection of heart diseases.

Early diagnosis of skin cancer, especially melanoma, is important in saving lives of patients. Despite the fact that deep learning has enhanced automated lesion classification, the current models are characterized by high variance, low domain robustness, and low interpretability. We introduce a explainable ensemble which is a dense net 121 plus inception v3 plus a dense net plus squeeze-and-excitation (SE) attention to learn both complementary channel and spatial features. Lesion-specific regions are visualized in Grad-CAM with the guidance of the predictions, which boosts clinical trust. The experiments on the HAM10000 dataset demonstrate that the ensemble has 97.41% accuracy, which is higher than the individual models and state-of-the-art methods. This framework offers a credible and clear solution to Computer-Aided Diagnosis (CAD) in dermatology.

Diabetes is a serious global health issue and early detection is important to avoid serious complications and ensure positive results for the patient. Currently, the standard process of diagnosing the condition is the use of expensive and timeconsuming invasive blood tests. However, this study aims to provide a new method of automatically classifying the stage of diabetic condition by utilizing the visual properties of the patient's tongue. The proposed method involves processing the images of the tongues by identifying the vital visual properties such as the color, coatings, and texture of the tongues that are correlated to the health of the metabolism of a patient. A combination model of a convolution neural network and a radial basis function neural network is designed to classify the stage of the diabetic condition of a patient. This model is trained on an expanded pool of 6,000 images that are initially obtained from 200 images of the tongues of patients with three levels of conditions (healthy, pre-diabetes, and diabetic). Experimental findings clearly establish the fact that the hybrid model has been able to produce a 95% level of overall accuracy with high precision, recall, and F1-score for all classes. Based on the trained model, an users-friendly web interface is designed to make predictions and offer personalized recommendations for lifestyle changes. Experimental findings clearly establish the fact that tongue image analysis can be a simple effective means for diagnosing diabetes, ranging from a noninvasive approach to the conventional ones. This approach can easily help make a significant contribution to the management as well as prevention of diabetes.

Deep face recognition has achieved near-saturated performance on several controlled benchmarks, yet its robustness in unconstrained real-world scenarios remains a critical challenge. This paper presents a comparative analysis of three backbone architectures,a classification-based SE-ResNet-50 and two metric-learning-based models, IR-ResNet-50 and IR-SE-ResNet-50 trained with ArcFace loss on the large-scale VGGFace2 dataset and evaluated on the challenging IJB-C verification benchmark. All models are trained from scratch on a 4,605-identity subset of VGGFace2, achieving strong in-domain performance with validation accuracies in the 94-96% range and verification AUC of approximately 0.955 on VGGFace2 validation pairs. However, when evaluated on the IJB-C 1:1 verification protocol, the same models exhibit a dramatic drop in performance. IR-ResNet-50 with ArcFace achieves only 1.10% true accept rate (TAR) at false accept rate (FAR) 10-3 and 0.36% at FAR 10-4. IR-SE-ResNet-50 with alignment and MTCNN-based cropping marginally improves TAR to 1.75% at FAR 10-3 and 0.57% at FAR 10-4, while SE-ResNet-50 trained with softmax classification surprisingly performs best among the three, reaching 28.02% TAR at FAR 10-3 and 11.99% at FAR 10-4. These results highlight a severe generalization gap between curated training data and unconstrained benchmarks and suggest that strong margin-based losses alone are insufficient to guarantee cross-domain robustness. We further analyze failure modes related to pose, alignment mismatch, and template noise, and motivate a Pose-Aware Feature Rectification Network (PAFR-Net) as a future direction to improve pose robustness by learning pose-conditioned residual corrections in feature space.

The rapid growth of e-commerce has made product recommendation systems essential for enhancing customer experience and driving business success. This research proposes an advanced recommendation framework that integrates sentiment analysis (SA) and collaborative filtering (CF) to improve recommendation accuracy and user satisfaction. The methodology involves feature-level sentiment analysis with a multi-step pipeline: data preprocessing, feature extraction using a log-term frequency-based modified inverse class frequency (LFMI) algorithm, and sentiment classification using a Multi-Layer Attention-based Encoder-Decoder Temporal Convolution Neural Network (MLA-EDTCNet). To address class imbalance issues, a Modified Conditional Generative Adversarial Network (MCGAN) generates balanced oversamples. Furthermore, the Ocotillo Optimization Algorithm (OcOA) fine-tunes the model parameters to ensure optimal performance by balancing exploration and exploitation during training. The integrated system predicts sentiment polarity—positive, negative, or neutral—and combines these insights with CF to provide personalized product recommendations. Extensive experiments conducted on an Amazon product dataset demonstrate that the proposed approach outperforms state-of-the-art models in accuracy, precision, recall, F1-score, and AUC. By leveraging SA and CF, the framework delivers recommendations tailored to user preferences while enhancing engagement and satisfaction. This research highlights the potential of hybrid deep learning techniques to address critical challenges in recommendation systems, including class imbalance and feature extraction, offering a robust solution for modern e-commerce platforms.

The propagation of IoT devices in healthcare, environmental monitoring, and smart cities has created the demand for robust anomaly detection mechanisms for security against failure and cyber-attacks. Most of these devices, however, must run within power-constrained environments, and energy efficiency takes the center stage. Conventional anomaly detection techniques, relying on power-intensive operations and periodic transmission of data, have the potential to drain IoT devices' short power supplies in no time, rendering them inappropriate for continuous use. To overcome this, this paper suggests a new energy-aware anomaly detection framework tailored to IoT systems with short power supplies. Our approach uses lightweight machine learning with adaptive sampling schemes to facilitate continuous monitoring without wasting power. The model makes use of in-device feature selection and edge computing models to lessen data transmission requirements and processing requirements, with a hierarchical detection framework where initial screening is done at the device level and only high-priority anomalies are passed on to edge processors for in-depth analysis. Results show significant energy saving without compromising accuracy, proving the applicability of the model to real-world IoT environments with power limitations.

Data analytics and data processing are continuously increasing in cloud applications. They should face the challenges of data scalability, smart, and stable workflow management. The existing data processing applications are not scalable and fault tolerant. Hence, this research aims to provide an adaptive workflow orchestration framework that can manage the entire workflow automatically, with improved scalability, and handle a large volume of data using a directed acyclic graph available in Apache Airflow. The proposed framework can handle multidimensional and multilevel data processing flow in any distributed environment. The framework monitors continuously in a real-time environment, and it uses effective and dynamic scheduling, dynamic prioritization, and confirms the data flow even in inconsistent workloads. To do this, a machine learning algorithm is implemented to predict the availability of requested resources, bottlenecks in resource allocation, and the execution path to ensure automatic data processing in the cloud. It also reduces the overall execution time and functional overhead problems. The machine learning algorithm learns the historical data efficiently to make accurate decisions to reduce pipeline letdowns. To demonstrate the overall process and performance of the framework, the Kubernetes, BigQuery, Amazon S3, and native services are integrated to provide improved reliability and scalability when deploying the framework processes in multi- and hybridcloud architectures. It also includes the checkpoints and recovery points that provide robust fault management in the entire architecture. From the experimental outputs, the proposed orchestration frameworks outperform the others in terms of throughput, latency reduction, scheduling, fault-tolerance, and cost. This research also highlights that the integration of automations, intelligence, and adaptability in the cloud workflow framework and its potential value can be used in the next generation of cloud data processing applications.

The large-scale production of multimodal fake news, combining text and images, presents significant detection challenges due to distribution discrepancies. Traditional detectors struggle with open-world scenarios, while Large Vision-Language Models (LVLMs) lack specificity in identifying local forgeries. Existing methods often overestimate public opinion’s impact, failing to curb misinformation at early stages. This study introduces a Modified Transformer (MT) model, fine-tuned in three stages using fabricated news articles. The model is further optimized using PSODO, a hybrid Particle Swarm Optimization and Dandelion Optimization algorithm, addressing limitations such as slow convergence and local optima entrapment. PSODO enhances search efficiency by integrating global and local search strategies. Experimental results on benchmark datasets demonstrate that the proposed approach significantly improves fake news detection accuracy. The model effectively captures distribution inconsistencies and multimodal forgery details, outperforming conventional detectors and LVLMs. This research highlights the importance of integrating transformers and hybrid optimization to develop generalized, scalable, and accurate fake news detection systems.

Face Recognition is one of the most advanced and drastically growing research areas because it helps identify people globally in various ethical and unethical applications. Face recognition needs face detection that can be compared with a list of available faces to predict the correct person. Face detection has become popular, easy, and fast since it follows the Viola-Jones FD method. Face comparison is obtained by comparing the internal and external information from the face images, like different features, face structure, key points, and patch-by-patch comparison. Earlier face recognition methods used separate algorithms for feature extraction from the face images, like color, shape, texture, histogram, and local and global binary pattern, to compare pairs of images where they provide more complexity regarding computation, cost, and time. After the evolution of artificial intelligence models, recent research has focused on using machine and deep learning algorithms for face detection and recognition. However, the accuracy of face recognition models needs to be improved under various conditions. Thus, this paper used a two-stage face comparison model to enhance face recognition efficiency. A consequence of three CNN models called CNN-1, CNN-2, and CNN-3 are used to detect the faces, detect the facial features, and recognize the faces, respectively. The CNN models are implemented in Python, and the results are verified by experimenting with multiple benchmark face datasets. The output accuracy obtained from the face detection and recognition is compared with the facial feature detection and recognition to choose the best to identify the criminals. From the comparison, both FDR and FFDR obtained 99.68% accuracy equally

The main objective of this paper is to implement a transfer learning model for predicting anomalies in online streaming data. Streaming data is a continuous data generation and transmission model with a huge amount of data, enabling different kinds of vulnerable attacks in the network. It leads to negative impacts on the overall network performance. Several earlier methods have been proposed to improve anomaly detection accuracy in streaming data, whereas the false positive rate is high. This paper has aimed to increase the anomaly detection rate with a reduced false positive rate. Hence, it proposed a novel transfer learning method for designing an effective anomaly detection model in data streaming applications. It implements a long., short-term memory for managing the continuous generation and transfer of data called streaming data because it has multiple built-in features like forget gate., which operates the memory by eliminating unwanted and redundant data flows in the streaming process. The LSTM model is deployed in a kind of MANET called VANET, where it is applied to detect anomalies during vehicle communication. This paper provides high prediction accuracy since it integrates various data analytics tasks, like preprocessing, feature extraction, and classification, which feed quality data and perform fast analysis. The LSTM can detect anomalies, including DoS, DDoS, Sybil, Sinkhole, Wormhole, and blackhole. The simulation is carried out by implementing LSTM in Python and executed on a benchmark dataset to verify the efficacy of LSTM. The output shows that the model provides higher accuracy, low latency, and high throughput and is suitable for many real-time applications like IoT networks and cybersecurity.

With the rapid growth of Internet technology and social networks, the generation of text-based information on the web is increased. To ease the Natural Language Processing (NLP) tasks, analyzing the sentiments behind the provided input text is highly important. To effectively analyze the polarities of sentiments (positive, negative and neutral), categorizing the aspects in the text is an essential task. Several existing studies have attempted to accurately classify aspects based on sentiments in text inputs. However, the existing methods attained limited performance because of reduced aspect coverage, inefficiency in handling ambiguous language, inappropriate feature extraction, lack of contextual understanding and overfitting issues. Thus, the proposed study intends to develop an effective word embedding scheme with a novel hybrid deep learning technique for performing aspect-based sentimental analysis in a social media text. Initially, the collected raw input text data are pre-processed to reduce the undesirable data by initiating tokenization, stemming, lemmatization, duplicate removal, stop words removal, empty sets removal and empty rows removal. The required information from the pre-processed text is extracted using three varied word-level embedding methods: Scored-Lexicon based Word2Vec, Glove modelling and Extended Bidirectional Encoder Representation from Transformers (E-BERT). After extracting sufficient features, the aspects are analyzed, and the exact sentimental polarities are classified through a novel Positional-Attention-based Bidirectional Deep Stacked AutoEncoder (PA_BiDSAE) model. In this proposed classification, the BiLSTM network is hybridized with a deep stacked autoencoder (DSAE) model to categorize sentiment. The experimental analysis is done by using Python software, and the proposed model is simulated with three publicly available datasets: SemEval Challenge 2014 (Restaurant), SemEval Challenge 2014 (Laptop) and SemEval Challenge 2015 (Restaurant). The performance analysis proves that the proposed hybrid deep learning model obtains improved classification performance in accuracy, precision, recall, specificity, F1 score and kappa measure.

Precision agriculture aims to improve agricultural productivity by combining technology and farming. The productivity of the cotton is mostly effected by leaf diseases, if we predict these diseases at early stage it helps the farmers to improve productivity. We put forward a novel method for cotton leaf disease detection that works on hybrid dataset which compromises of images from the kaggle dataset and real-time images. Deep learning models VGG16 and VGG19 are applied on this hybrid dataset for disease detection of cotton leaves. This research will not only contribute to improve crop health but also be a valuable resource for farmers. A systematic comparison of the VGG16 and VGG19 models reveals their functional differences in disease detection. VGG16 and VGG19 has achieved accuracy of 94% and 95% in disease detection

In agriculture, pests are the major reason that causes low yield and greatly affects the crop. Cotton plays a major role in the textile industry and due to a lack of pest identification more amount of cotton crops is getting damaged. To solve this problem, Convolutional neural networks along with MobilenetV2 is used to detect the pests in the cotton plant by passing an RGB image to the proposed model and the model detects whether the pest is present in the crop or not with the accuracy rate of 96.31%. If the pest is present, then the farmer has to be ready with pesticide and if the pest is not present, the farmer has to give fertilizer based on soil properties. This can be done by using the Random Forest (RF) algorithm with an accuracy of 97.95%. This can help farmers to produce more yield.

Requirement engineering’s challenges become manageable when applied to the global advancement of programming. There are numerous reasons why something is difficult. Chances could be one of them since the global improvement perspective is more open to gambling. Therefore, it could be one of the main justifications for taking requirement engineering testing seriously. To begin with, it is necessary to identify the factors that genuinely result in these threats. This essay then separates the factors as well as the risks that these elements may bring about. In the context of the global programming improvement viewpoint, an orderly writing survey is completed for the observable evidence of these variables and the risks that may occur during the necessity designing cycle. The list suggests a moderate improvement in aiding exercises in necessity designing in a global programming advancement worldview. This work is very beneficial for those with less experience working in global programming advancement.

Ad hoc network nodes are aggregate data packet from different environment; there is multiple path communication causing the sudden energy depletion in network. This type of energy loss can lead to failure of connectivity between the two intermediate nodes. If link gets failure, then it has frequent loss of data packets. Less energy nodes do not classify data from the network structure. It reduces packet delivery ratio and increases the energy consumption. The proposed cross-layer method for data agglomeration (CLA) is designed to organize the data packet frequently among the various communication routes; the nodes in the path can able to proceed packet organization for the support of cross-layer scheme. Magnificent path discovery algorithm is constructed to offer the better packet collection route to target node. This process uses multisource node with multiple path for packet transmission in network. It minimizes the energy consumption and increases the packet delivery ratio. The simulation parameters are delay, detection efficiency, energy consumption, network lifetime, and packet delivery ratio.

Underwater mining of minerals and rocks is a highly challenging task before the discovery of SONAR (Sound Navigation and Ranging) system. Lately, the mine detection process was performed by the divers trained in the disposal of hazardous ordnance, marine mammals, video cameras mounted on mine-neutralization trucks, and laser systems. which leads to risk and loss to the marine life. SONAR system is capable of capturing Scan-side sonar images, but the model's accuracy is a concern. So Naval defense system need to use a much more accurate system as mines can be easily mistaken as rock, to obtain accurate results we will be working on the dataset of frequencies. Recently, this prediction system was constructed using many machine learning methodologies. This research study proposes to apply XGBoost algorithm to develop a prediction system to predict whether the object is rock or mine. Here, the accuracy of the proposed model is compared with the accuracy of the existing models.

Identification of fresh air by predicting air quality Index is very important for providing better healthy environment to the society. Air pollution causes a severe health issues for the humans as well as threat to the environment. Air quality is measured by predicting air quality Index using some parameters. Based on air quality index value range it'll help to forecast the levels of human health concerns. This study proposes XGboost algorithm to forecast the air quality. When compared to other machine learning models, XGboost helps to predict the air quality with high accuracy rate.

The major contribution to Indian economy comes from agriculture which stands as the backbone and also it is the livelihood of many farmers. But now-a-days the crops are being infected by multiple diseases and causing widespread of the disease which in turn damages the entire crop fields if they are not noticed in prior. Crops get diseased by fungi, virus and bacteria and also by worms and insects that attack the crops. These crop diseases should be diagnosed with the help of emerging technologies like DL (Deep Learning) which provide plenty of techniques for disease detection. The proposed model which uses a Super Resolution Convolutional Neural Network (SRCNN) to improve the quality of the crop leaf images and also a CNN which acts as a classifier that helps to detect the crop disease. When the model is trained with SRCNN it improves the performance and illness of crops is also identified in an unerring way.

In the current years, the Home Automation organizations takes to see a rapid changes due to introduction of many wireless technologies. The detonation in the wireless expertise has gotten the arrival of countless ethics, particularly in the ISM (engineering, science, and medicine) receiver band. Wi-Fi is an IEEE 802.11 standard. Protocol household usual for data roads with business and consumer devices. Wi-fi is besieged at bids that requires high data rate, wide-ranging mobile life, and locked interacting. Wi-fi has a defined rate of 11 Mbit/s (Megabits per subsequent), best suitable for periodic or discontinuous data from a sensor or input device, or a single signal driver. Wireless home automation systems are meant to be installed in existing homes atmospheres, lacking any changes in the infrastructure. The reconstruction centers on gratitude of voice orders and uses dynamic wi-fi wireless communiqué modules along with microcontroller. This plan is most fit for the elderly in addition the disabled persons specially those who live unaided and since find voice, so it is secure. The home-based computerization system is projected to control all lights and electrical engagements in a home or place of work using voice guidelines. So, in this communication, our aim is to calculate a voice gratitude wireless wi-fi built home computerization. To achieve this, the proposed research work has used ESP 8266 module along with Alexa and Google assistant to decrease power consumption and enhance the security.

In the advanced innovative present reality, most of the public are depending on different surveys for different utilizations and different products. Thus, to examine these audits and comprehend whether they are supporting or refuting about, opinion investigation can be utilized. The days for the remark surveys are currently disappearing and everybody is keen on hearing the audits so as they need not stop their functions. Thus, a slant investigation that does both through brief snippet and text can be used to complete work in sound. Along these lines, in this exploration paper, different Machine Learning and Deep Learning Models like Naive Bayes, S upport Vector Machine (S VM), Random Forest, and Multi-Layer Perceptron are used. Additionally, as of to change over voice to a message, a google API and a deep learning technique are used and best of two is exhibited and afterward sentiment analysis on those texts is performed and finally the sentiment (i.e., positive, or negative) of the speech is obtained. All these processes are carried out in real-time.

A Classification method stands out as a reliable data mining technique applied in medical sciences. We observe multi-classification problem afflicting many recent applications that includes social network analysis, biology, anomaly detection and computer vision. However, such classification techniques usually struggle while dealing with features of data that is generated from multiple classes. Moreover, in case of large medical data, it is observed that the dimension of the data poses the biggest challenge while applying classification technique to them. In order to overcome such problems, we have proposed an adaptive random forest classifier method that uses ensemble feature selection technique for better information gain (IG), improved correlation (IC) and gain ratio (GR). Also, it seeks to solve the class imbalance problem by applying bootstrap resampling for medical data. The result of the proposed method proved that adaptive RF (Random Forest) classifier offers better accuracy, precision and F-score values than standard Random Forest and KNN classification algorithms. The overall performance of algorithms was tested over five real datasets. The result analysis shows performance of the proposed classifier is promising in all real datasets as compared to standard methods.

The success of a project depends on accurate effort estimations, managers are always under pressure to prepare accurate effort estimations, in COCOMO model to estimate the effort it requires project parameters. Identification of the type of the project and choosing the project parameters are very important aspect, accurate project parameters generation and estimations are coming from mature organizations others owing to lack of history databases. Estimations are based on lines of code (size of the project), functionality of the project. If the project estimations are based on size of the project the Constructive Cost Model plays vital role. This work explains an expert system that integrates conventional and Soft Computing techniques for dealing uncertainty i.e. virtual nodes generated in the decision paths at leaf node level and we will create an additional node, the main objective of the paper is how to generate suitable values for these nodes. For this purpose we propose the method to generate project parameters using fuzzy logic. In this work solid line indicates the project already done; dotted line indicates the project with uncertainties. After the project parameters are generated using Fuzzy Logic then effort estimations can be prepared.

Every educational system organizational goal is to provide a good and fruitful knowledge to the students. Now a days most of the educational institutions are spending most of their time and economy on finding out students' performance. By analyzing the performance, they identify certain cluster of the students for whom they must give extra bit of care and actions, so that they performance gets enhanced. Researchers have recently proposed several machine learning-based algorithms for predicting academic achievement. In this paper, Linear Regression algorithm and Random Forest algorithm are used to predict a student's academic achievement. On the basis of confusion matrix, accuracy, precision, recall, and F1 ranking, the performance of two algorithms was compared to that of existing algorithms. The Random Forest algorithm-based prediction is more accurate, according to the results report.

The major concern of WSN based application is life time of sensor nodes. Clustering is one of the energy efficient techniques for improving life time of WSN and mainly reduces communication overhead between nodes. Selecting an efficient CH plays imperative role in extending the lifetime of WSN. The trust concept plays an important role in WSN and it’s generally used for detecting malicious, selfish and faulty nodes in a network which increase communication overhead and energy consumption. To conquer these issues, trust based cluster head selection and routing method is proposed. A trust level is computed for each node based on communication, energy and neighbor nodes. This paper proposes an efficient method for cluster head selection and secures routing using two evolutionary algorithms. Cat Swarm Optimization (CSO) is used to select the cluster heads and Firefly algorithm is used for secure routing.Extensive simulations are conducted on various circumstances. The simulation results shows that the proposed trust based CSO finds the optimal cluster head and firefly algorithm discovers the optimal paths which improves the network lifetime and reduces end-to-end delay compared to other techniques.

Wireless Sensor Network (WSN) is a system with huge number of sensors connected to one another by placing them in a specific area. Different issues with WSN includes (but not limited to) the coverage, network lifetime and aggregation. The lifetime of a network can be improved by the clustering with the reduction of energy consumption. Clustering will group the related type of sensors into a single place with a head sensor node for message aggregation and transmission between other nodes and Base Station (BS). The cluster head (CH) consume more energy, when aggregating and transmitting the data. With the suitable identification of CH, there will be a reduction in the consumption of energy and improves the life of Wireless Sensor Network to be more. This paper modifies the meta-heuristic algorithms for improving the network lifetime by choosing appropriate cluster head and optimal path. K-Genetic Algorithm (K-GA) is proposed for efficient cluster head selection. Initially, the sensors are clustered using k-means clustering based on their location and Genetic Algorithm has been applied to detect the best cluster head. For secure optimal routing, Trust based Firefly (T-FA) path selection algorithm is used. Extensive simulations are conducted on various circumstances. The results obtained on the simulation indicates that the proposed K-GA helps in determining the optimized head of the cluster and T-FA discovers the optimal paths which enriches the life of the network by reducing end-to-end delay compared to other techniques.

Wireless sensor networks (WSNs) contain an excellent quantity of battery-driven small nodes that have sensing, computing and communication capabilities. Therefore, it's essential to design an energy-efficient routing protocol to cleverly use the limited energy of WSNs. One-way of managing the energy efficiency is grouping sensors to form a cluster and choose a node as lead to manage referred to as cluster head (CH). In case a malicious node or lower energy node is chosen as a cluster head, the throughput of the network is greatly affected. Thus selection of cluster heads with higher trust and residual energy becomes crucial for the overall network performance. To deal with this problem, this work proposes a Trust based Cluster Head Selection with Secure Routing (TCHS_SR) algorithm for wireless sensor network. The proposed method relies on an effective distributed trust model for cluster head selection and it also considers the secure route for data transfer. The experimental result shows that the proposed TCHS_SR algorithm reduces energy consumption and end-to-end delay, furthermore increases the throughput and packet delivery ratio efficiently.