The exponential growth of digital technology and the widespread use of web applications have led to an overwhelming increase in data, necessitating effective recommender systems to filter and present relevant content. Recommender systems have evolved as essential tools in the digital era, providing personalized suggestions by analyzing user interests and behavior patterns. Integration of these systems has become an essential component of digital marketing strategies. They hold significant positions in a wide range of domains, including streaming services (music, movies, and books), social media systems, digital governance, electronic commerce, e-libraries, e-learning systems, traveling and resource services, and many more. Recently, recommender systems have been extensively employed in the healthcare and educational sectors to assist users in identifying and accessing content that aligns with their interests. While these systems offer numerous advantages, they are also confronted with several formidable challenges, including the cold-start problem, data sparsity, the presence of gray sheep users, starvation, and shilling, which can negatively impact their performance. Extensive research has been conducted to address these challenges and improve the accuracy of recommender systems. This review provides a comprehensive overview of the primary methods, including Content-Based Filtering, Collaborative Filtering, Knowledge-Based Systems, Demographic-Based Systems, Community-Based Systems, Hybrid Systems, Cross Domain Systems, and Social Recommender Systems, along with their respective merits and demerits. It also examines the evaluation methods and metrics used to assess recommender systems performance. Additionally, the review explores the application of recommender systems across diverse domains and highlights the tools and technologies supporting their development. Future opportunities and unresolved issues in recommender systems are explored to help promote continued research and innovation.

In the era of digital technology, when material created by users is prevalent on online platforms, considerable difficulty is faced in analyzing large volumes of text in order to comprehend user emotions and forecast product ratings. The rapid rise in online reviews and comments necessitates the use of advanced tools to assess this data and extract valuable insights. This is considered crucial for the effectiveness of recommendation systems in many industries. This paper introduces the Review Text Granularity (RTG) Model, a new way to use the complex information in review texts to improve sentiment analysis and rating prediction. The RTG Model uses an advanced approach to scoring sentiments. It measures the strength of sentiments and gives a continuous sentiment score instead of a simple positive or negative label. This makes it different from other binary sentiment analysis methods. Multiple predictive modeling techniques are used, which makes it possible for this comprehensive sentiment analysis to greatly improve the accuracy of rating predictions. It has been shown by research that the depth of textual reviews is better captured and measured by the RTG Model, providing a more detailed and precise picture of user opinions. The RTG Model works really well, making rating predictions in recommendation systems more accurate and useful. A detailed study using a real-world dataset of IMDb movie reviews demonstrated this. The study emphasizes the benefits of utilizing intricate sentiment scores in addition to conventional rating data. The potential applicability of the RTG Model in several fields, including entertainment, e-commerce, and social media, is demonstrated, leading to enhanced and tailored user experiences.

Among all types of cancer, one of the most serious types of cancer that causes the unpredictability of mortality in the majority of women across the world is breast cancer (BC). Numerous researches have been done for the BC treatment and also are going on at present at utmost rate which leads to the better understanding of the disease and hence provides better results. Thus accurate prediction of BC plays a vital role in efficient treatment and also for its early stage detection. Since a few years ago, many Machine Learning (ML) algorithms have been suggested for precise BC prediction. In order to improve BC prediction using machine learning analytics, this article also suggests the MLESBCD (Machine Learning Expert System for Breast Cancer Detection). Prior to using the classification algorithm, the MLESBCD the BC data is preprocessed to handle outliers and the missing values. On the BC dataset from the UCI machine learning repository, the suggested MLESBCD generates satisfactory accuracy.

The exponential growth of online multimedia content across various platforms has created an urgent need for robust assistive technologies to manage the overwhelming volume of information. Consequently, considerable efforts have been dedicated to developing sophisticated multimedia recommendation systems, with Neural Collaborative Filtering (NCF) emerging as a prevalent methodology. However, the conventional NCF method exhibits significant limitations, particularly in integrating contextual data and effectively handling sparse and imbalanced datasets. To address these limitations, this research introduces the Contextual Neural Collaborative Filtering (C-NCF) method, which enhances the NCF framework by incorporating contextual data to enrich the learning process of user-item interactions. The primary objective of this method is to improve rating prediction accuracy, a crucial factor in generating more effective recommendations. A key innovation of the C-NCF method lies in its interaction mechanism, where user ratings of items are evaluated under diverse contextual conditions, assigning varying importance to each contextual factor. Extensive testing on three real-world datasets demonstrated that the C-NCF method outperforms existing advanced techniques. Empirical findings demonstrate that the C-NCF method achieved an average error reduction of 36.43% in Mean Absolute Error and 36.60% in Root Mean Squared Error compared to traditional collaborative filtering, matrix factorization, and context-aware models, significantly enhancing recommendation quality. These insights open promising avenues for further exploration in the field of context-aware recommender systems.

Data mining is the most widely used method for discovering knowledge. There are numerous data mining tasks, with classification being the most frequently encountered task in various application domains such as fraud detection, disease diagnosis, text classification, and so on. Many classification techniques, such as Bayesian classifiers, decision trees, genetic algorithms, neural networks (NNs), and so on, are available to help researchers solve problems in a variety of domains. However, NNs are the most frequently used classification approach because they are effective at solving classification problems that cannot be divided into linear and non-linear categories, have high classification accuracy on large datasets, and require minimal processing effort. Despite having good classification performances, NNs have a pitfall associated with them which hinders their applicability in some real-world applications. NNs are black boxes in nature, which means they cannot make transparent decisions that humans can interpret. Because of this limitation, NNs are unsuitable for many applications that require transparency in decision-making as well as high accuracy, such as audit mining or medical diagnosis. The well-known solution to this inherent disadvantage of NNs is to extract explainable decision rules from them. The extracted rules provide a detailed understanding of how NNs work in a human-readable format. Rule extraction is a well-established technique with a plethora of literature on the subject. However, there are very few papers whose primary goal is to survey the existing literature. As a result, the goal of this work is to provide a detailed analysis of the existing literature and to create a framework for existing and new researchers to conduct research in this field. The paper examines the state-of art from the perspective of designing framework of the algorithms, evaluation criteria, and applications.

In data mining and machine learning communities, Neural Network (NN) is a popular classification method. On extremely unbalanced and complicated datasets, NN may achieve excellent classification accuracy. However, one disadvantage of NN is its inability to explain its reasoning process, which restricts its use in numerous sectors that need clear conclusions as well as high accuracy. To address this issue, rule-extraction mechanisms exist that extract intelligible classification-rules from NN and turn them into a white box. Attribute or network pruning, dealing with diverse attribute types, rule pruning, and dealing with class overlapping difficulties are all significant components or portions of many existing rule extraction methods, and present strategies to deal with these aspects are insufficiently successful. As a result, this study offers a rule extraction approach named “Comprehensible and Transparent Rule Extraction Using Neural Network”-CTRENN to address the aforementioned shortcomings and transform NN into a white box with high accuracy and better explain-ability. The suggested CTRENN is an expansion of the state-of-art Rule Extraction from Neural Network Using Classified and Misclassified Data technique (RxNCM). The CTRENN augments the RxNCM with a floating sequential search for feature and rule selection to improve feature and rule selection. CTRENN also distinguishes between continuous and discrete properties to improve the readability of the produced rules. Unlike RxNCM, the CTRENN employs a probabilistic technique to deal with the overlapping of attribute data ranges in various classes. Experiments are carried out using six real life datasets obtained from the UCI repository in order to illustrate the efficacy of the proposed CTRENN algorithm in comparison to the current methods.

The accurate prediction of heart disease is crucial in the field of medicine. While convolutional neural networks have shown remarkable precision in heart disease prediction, they are often perceived as opaque models due to their complex internal workings. This paper introduces a novel method, named Extraction of Classification Rules from Convolutional Neural Network (ECRCNN), aimed at extracting rules from convolutional neural networks to enhance interpretability in heart disease prediction. The ECRCNN algorithm analyses updated kernels to derive understandable rules from convolutional neural networks, providing valuable insights into the contributing factors of heart disease. The algorithm’s performance is assessed using the Statlog (Heart) dataset from the University of California, Irvine’s repository. Experimental results underscore the effectiveness of the ECRCNN algorithm in predicting heart disease and extracting meaningful rules. The extracted rules can assist healthcare professionals in making precise diagnoses and formulating targeted treatment plans. In summary, the proposed method bridges the gap between the high accuracy of convolutional neural networks and the interpretability necessary for informed decision-making in heart disease prediction.

Symbolically representing the knowledge acquired by a neural network is a profound endeavor aimed at illuminating the latent information embedded within the network. The literature offers a multitude of algorithms dedicated to extracting symbolic classification rules from neural networks. While some excel in producing highly accurate rules, others specialize in generating rules that are easily comprehensible. Nevertheless, only a scant few algorithms manage to strike a harmonious balance between comprehensibility and accuracy. One such exemplary technique is the Rule Extraction from Neural Network Using Classified and Misclassified Data (RxNCM) algorithm, which adeptly generates straightforward and precise rules outlining input data ranges with commendable accuracy. This article endeavors to enhance the classification performance of the RxNCM algorithm by leveraging ensemble technique. Ensembles, a burgeoning field, focus on augmenting classifier performance by harnessing the strengths of individual classifiers. Extraction of rules through neural network ensembles is relatively underexplored, this paper bridges the gap by introducing the Rule extraction using Neural Network Ensembles (RENNE) algorithm. RENNE is designed to refine the classification rules derived from the RxNCM algorithm through ensemble strategy. Specifically, RENNE leverages patterns correctly predicted by an ensemble of neural networks during the rule generation process. The efficacy of the algorithm is validated using seven datasets sourced from the UCI repository. The outcomes indicate that the proposed RENNE algorithm outperforms the RxNCM algorithm in terms of performance.

Parkinson’s disease (PD) is a common progressive neurodegenerative disorder that occurs due to corrosion of the substantianigra, located in the thalamic region of the human brain, and is responsible for the transmission of neural signals throughout the human body using brain chemical, termed as “dopamine.” Diagnosis of PD is difficult, as it is often affected by the characteristics of the medical data of the patients, which include the presence of various indicators, imbalance cases of patients’ data records, similar cases of healthy/affected persons, etc. Hence, sometimes the process of diagnosis may also be affected by human error. To overcome this problem some intelligent models have been proposed; however, most of them are single classifier-based models and due to this these models cannot handle noisy and imbalanced data properly and thus sometimes overfit the model. To reduce bias and variance, and to avoid overfitting of a single classifier-based model, this paper proposes an ensemble-based PD diagnosis model, named Ensembled Expert System for Diagnosis of Parkinson’s Disease (EESDPD) with relevant features and a simple stacking ensemble technique. The proposed EESDPD aggregates diverse assumptions for making the prediction. The performance of the proposed EESDPD is compared with the performances of logistic regression, SVM, Naïve Bayes, Random Forest, XGBoost, simple Decision Tree, B-TDS-PD and B-TESM-PD in terms of classification accuracy, precision, recall and F1-score measures.

Monitoring systems are becoming more crucial for city security, specifically to address any kind of violence. Manual detection of violence through video surveillance requires well-trained human resources. However, accuracy and speed are compromised in the case of manual approach. The lack of human resources and their limited ability to interpret violence from a video is the major concern of this approach. Thus, automatic violence detection is required to take timely actions. Most of the existing work for violence detection have used either traditional Machine Learning (ML) based approach or a combination of Convolution Neural Network (CNN) and Long Short Term Memory (LSTM) network for incorporating Spatio-temporal relations. However, video clips contain an extensive number of frames which contain unnecessary and redundant information. Elimination of such frames increases the efficiency and accuracy of the system. Thus, an Efficient ConvLSTM based Violence Detection System (ECLVDS) model is proposed to address this issue. The proposed model uses a clustering-based keyframe extraction technique to eliminate redundant and useless frames. Moreover, the proposed model uses a ConvLSTM model instead of a simple LSTM because the ConvLSTM network extracts better spatiotemporal correlation than the normal LSTM model. The proposed model is evaluated with the widely used benchmark Hockey Fight Dataset and achieved an accuracy of 98.90%, which is comparatively better than some of the existing models.

Rainfall prediction is a significant and demanding task in today's world. Due to the nonlinearity and complication, rainfall and climate require advanced computer modeling and simulation to predict accurately. This paper comes up with a comprehensive survey and a fair comparison of several types of neural network architectures for rainfall prediction and mentions the challenges of using traditional and advanced methods in doing this task. Finally, the paper discusses the utility of deep learning methods for rainfall forecasting.

Neural networks are well-known for their impressive classification performance, and the ensemble learning technique acts as a catalyst to improve this performance even further by integrating multiple networks. However, neural network ensembles, like neural networks, are regarded as a black box because they cannot explain their decision-making process. As a result, despite their high classification performance, neural networks and their ensembles are unsuitable for some applications that require explainable decisions. However, the rule extraction technique can overcome this drawback by representing the knowledge learned by a neural network in the guise of interpretable decision rules. A rule extraction algorithm provides neural networks the ability to justify their classification responses using explainable classification rules. There are several rule extraction algorithms for extracting classification rules from neural networks, but only a few of them use neural network ensembles to generate rules. As a result, this paper proposes a rule extraction algorithm called Rule Extraction Using Ensemble of Neural Network Ensembles (RE-E-NNES) to demonstrate the high performance of neural network ensembles. RE-E-NNES extracts classification rules by ensembling several neural network ensembles. The results demonstrate the efficacy of the proposed RE-E-NNES algorithm in comparison to other existing rule extraction algorithms.

Mortality rate due to fatal heart disease (HD) or cardiovascular disease (CVD) has increased drastically over the world in recent decades. HD is a very hazardous problem prevailing among people which is treatable if detected early. But in most of the cases, the disease is not diagnosed until it becomes severe. Hence, it is requisite to develop an effective system which can accurately diagnosis HD and provide a concise description for the underlying causes [risk factors (RFs)] of the disease, so that in future HD can be controlled only by managing the primary RFs. Recently, researchers are using various machine learning algorithms for HD diagnosis, and neural network (NN) is one among them which has attracted tons of people because of its high performance. But the main obstacle with a NN is its black-box nature, i.e., its incapability in explaining the decisions. So, as a solution to this pitfall, the rule extraction algorithms can be very effective as they can extract explainable decision rules from NNs with high prediction accuracies. Many neural-based rule extraction algorithms have been applied successfully in various medical diagnosis problems. This study assesses the performance of rule extraction algorithms for HD diagnosis, particularly those that construct rules recursively from NNs. Because they subdivide a rule's subspace until the accuracy improves, recursive algorithms are known for delivering interpretable decisions with high accuracy. The recursive rule extraction algorithms' efficacy in HD diagnosis is demonstrated by the results. Along with the significant data ranges for the primary RFs, a maximum accuracy of 82.59% is attained.

Sign language (SL) is an alternative way of communication for hearing/speech impaired people. SL varies from country to country. Thus, different SLs came into existence such as American Sign Language (ASL), Chinese Sign Language (CSL), British Sign Language (BSL) and so on. In India, people use Indian Sign Language (ISL) as an alternative communication medium. Approximately, 2.7 million Indian people are deaf/dumb, out of which 98% population uses ISL for communication. But human interpreters are not available or affordable for identifying ISL. Thus, researchers are focussing on Sign Language Recognition System (SLRS) for the identification of ISL. But designing an SLRS for ISL is very difficult in comparison with other SL. ISL is very complex as it consists of single- and double-handed gesture and has an extensive vocabulary with similar gestures. This paper aims to provide a thorough study of ISL, its syntax/vocabulary and different trending techniques for designing an ISL recognition system. This paper also discusses various feature extraction techniques, classification methods used in ISL recognition and points out some challenges researchers face while developing an ISL recognition system.

A lot of people die all over the world due to various cardiovascular diseases. These diseases cause serious threat to the life of people and mostly affect people in the age range of 25-69years. There are various conventional methods and manual medical procedures of diagnosis for detection of cardiovascular diseases. However, these procedural methods are highly expensive and have reduced scalability performance, and hence are highly time consuming and sometimes prone to errors. Therefore, it is important to detect cardiovascular diseases using certain intelligent systems which deployMachine Learning (ML)techniques for better feature extraction, and early detection to avoid critical health circumstances, and thereby assist in improving medical diagnosis. This paper proposes and implements a cost effective and affordable ML model, which can detect whether a given person suffers from heart disease or not, given their clinical record, at an early stage. The proposed model is a composition of various ML classifiers such as K-Nearest Neighbors (KNN), Random Forest (RF), Decision Tree (DT), Logistic Regression(LR), Support Vector Machine(SVM) and Naïve Bayes(NB), on theCleveland dataset and makes a comparison amongst them. The model deploys data preprocessing such as normalization, feature extraction and classification. In this paper it is observed that the Naïve Bayes classifier gives best prediction which is 86.88%compared to all other classifiers and the Decision Tree classifier gives the least performance amongst all, which is 78.69%.

Everyone wishes to buy and live in a house which suits their lifestyle and which provides amenities according to their needs. There are many factors that are to be taken into consideration like area, location, view etc. for prediction of house price. It is very difficult to predict house price as it is constantly changing and quite often the prices are exaggerated for which people who want to buy houses, and various real estate agencies who want to invest in properties, find it difficult to buy or sell houses. For this reason, in this paper the author creates an advanced automated Machine Learning model using Simple Linear Regression, Polynomial Regression, Ridge Regression and Lasso Regression using the Boston house dataset to predict house price in future accurately, and to measure the accuracy of these models various measuring metrics like R-Squared, Root Mean Square Error (RMSE) and Cross-Validation are used. This paper also studies the correlation of various attributes of the Boston dataset using the heat map to see which attributes actually impact the prediction of the models. It removes the outliers which are present in the dataset to achieve good accuracy. In this paper it is observed that Lasso Regression performs better in all the measuring matrices whereas Simple Linear Regression performs poor in all of the measuring matrices.

Dengue Fever (DF) is one of the most severe epidemiologic disease which is caused by blood-feeding mosquitoes and also most speedily increasing public health issues in tropical and subtropical regions which sometimes cause continual epidemics. The general symptoms of dengue are fever which includes joint pain, headache, vomiting, etc.; however, if Dengue is not detected at an early stage and experience severe bleeding and shock, which may even lead to death. Thus, increasing DF can be very severe and critical, turning out to be a global threat. So, considering the severity of DF, this paper proposes a model, named Dengue Fever Expert System using Machine Learning Analytics (DFES-MLA) to explore the dengue data to predict the disease more efficiently considering only the symptomatic features. The dengue dataset is imbalanced in nature. And therefore, the DFES-MLA model uses the 4 most prominent oversampling techniques namely SMOTE, Borderline SMOTE, ADASYN, Gaussian SMOTE to get rid of the class imbalance problem of the dengue dataset. DFES-MLA model uses Decision Tree (DT) and Random Forest (RF) classifiers along with all the data pre-processing steps for DF prediction.

India's government spends 1.5 percent of its annual GDP on public healthcare, which is significantly less than that of other countries. Global public health spending, on the other hand, has almost doubled in line with inflation in the last two decades, reaching US $8.5 trillion in 2019, or 9.8% of global GDP. Multinational multi-private sectors provide around 60% of comprehensive medical treatments and 70% of out-patient care, which charge patients astronomically high fees. Because of the rising expense of quality healthcare, increased life expectancy, and the epidemiological shift toward non-communicable diseases, health insurance is becoming an essential commodity for everyone. Insurance data has increased dramatically in the last decade, and carriers now have access to it. The health insurance system explores predictive modeling to boost its business operations and services. Computer algorithms and Machine Learning (ML) is used to study and analyze the past insurance data and predict new output values based on trends in customer behavior, insurance policies, and data-driven business decisions, and support in formulating new schemes. Additionally, ML has found enormous and potential applications in the insurance industry. Thus, this paper develops a real-time insurance cost price prediction system named ML Health Insurance Prediction System (MLHIPS) using ML algorithms which will aid the insurance companies in the market for easy and rapid determination of values of premiums and thereby curb down health expenditure. The proposed model incorporates and demonstrates different models of regression such as Ridge Regression, Lasso Regression, Simple Linear Regression, Multiple Linear Regression and Polynomial Regression to anticipate insurance costs and assess model outcomes. In the proposed model, the Polynomial Regression model has achieved better results with an RMSE value of 5100.53 and R -squared value of 0.80 compared to all the other models.

Classification is one of the foremost machine learning tasks in this modern era. Neural Network (NN) is one of the powerful classification techniques. NN can achieve high classification accuracy on highly imbalanced and complex datasets, but lacks in explanation of its reasoning process which limits its applicability in various domains which require transparent decision along with good accuracy. There are some techniques which extract rules from NN and make it transparent; however, attribute pruning, rule pruning and class overlap algorithms are not sufficiently effective. Therefore, this paper proposes a rule extraction algorithm, called Transparent Rule Extraction using Neural Network (TRENN) to convert NN into white box with greater emphasis on attribute pruning and rule pruning. The proposed TRENN is a pedagogical approach and an extension of one of the existing algorithms named Rule Extraction from Neural Network using Classified and Misclassified data (RxNCM). The proposed TRENN extends the RxNCM with sequential floating backward search for feature and rule selection to improve the comprehensibility of the generated rules. Besides, the proposed TRENN uses probabilistic approach for the treatment of class overlapping problem in the rule updating phase instead of reclassification used in RxNCM where the overlap may persist. Experiments are conducted with eight real datasets collected from the UCI repository. Performance of the TRENN is measured with Precision, Recall, FP-Rate, F-measure, and local and global comprehensibility. It is observed from the experimental results that TRENN performs better than Re-RX, RxNCM and RxREN.

Sign Language (SL) is a parallel communication medium to spoken language, mostly used by people who are not able to speak or hear. A human interpreter is also required to work as a mediator or to act like a bridge to reduce communication gap between deaf/dumb people and normal people. However a well-trained human interpreter is not easily available and is also not cost-effective. Besides, it also compromises confidentiality and accuracy. To overcome these problems Sign Language Recognition System (SLRS) has been developed as an automatic and intelligent mediator for the communication. For a decade different Machine Learning (ML) techniques have been being used to design SLRS. This paper highlights about SL, gives a general framework of SLRS and reports different ML approaches proposed for the design of SLRS. This paper also reports Deep Learning (DL) models proposed for SLRS and finally highlights the challenges encountered in the design of SLRS.

Modern busy lifestyles are acting as a catalyst to enhance the growth of various health-related issues among people. As a consequence, a massive amount of medical data are getting accumulated every day. So, it is becoming a challenging task for the medical community to handle those data. In such a situation, if a system exists that can effectively analyze those data and can retrieve the primary causes of a disease, then the disease can be prevented on time by taking the correct precautionary measures beforehand. Recently, machine learning algorithms have been receiving a lot of appreciation in building such an expert system, and the neural network is one of them which has attracted a lot of researchers due to its high performance. But the main obstacle which hinders the application of neural networks in the medical domain is its black-box nature, i.e. its incapability in making a transparent decision. So, as a solution to this problem, the rule extraction process is becoming very popular as it can extract comprehensible rules from neural networks with high accuracy. Many rule extraction algorithms exist in the literature, but this paper mainly assesses the performances of the algorithms that generate rules recursively from neural networks. Recursive algorithms recursively subdivide the subspace of a rule until the accuracy increases. So, they can provide comprehensible decisions along with high accuracy. Four medical datasets are collected from the UCI repository for assessing the performances of the algorithms in diagnosing a disease. Results prove the effectiveness of the recursive rule extraction algorithms in medical diagnosis.

Nowadays, people are facing various health-related problems due to the modern life style what they follow. Breast Cancer is one of the most common problems among women worldwide which affects approximately 2.1 million women each year. Hence, it has become paramount to develop a system that can identify the major risk factors of Breast Cancer beforehand to make women aware about the risk factors and to take some precautionary measures to manage Breast Cancer. Consequently, this paper proposes a system called Transparent Breast Cancer Management System using P-Rules (TBCMS-PR) which identifies the major risk factors responsible for Breast Cancer in detail using decision tree and neural Network. TBCMS-PR uses decision tree to generate the rules for deciding the decision of Breast Cancer. Neural Network is used to keep only the relevant rules for Breast Cancer and to drop the irrelevant ones. Finally, the major risk factors with ranges are identified based on Sequential Search algorithm. The performance of the TBCMS-PR system is validated with the Breast Cancer dataset collected from UCI repository and is compared with a recent existing system. From the experimental results, it is observed that the proposed TBCMS-PR is significant and potential to manage Breast Cancer to a great extent by managing only one or two major risk factor(s).

Now a days Breast Cancer (BC) is a very common cancer among women and rarely man. BC incidence rate keeps increasing since 2000. Many intelligent systems have been proposed using different Machine Learning (ML) Techniques for early diagnosis of BC. Neural Networks (NN) is considered as one of the best ML classification techniques. In this paper we propose a medical expert system which named Transparent Neural Expert System for Breast Cancer (TNESBC) to manage BC by identifying Major Risk factors. The proposed TNESBC adopts the white-box NN model which named 'Rule Extraction from Artificial Neural Network(ANN) applying Classified and Misclassified data' (RxNCM) for rule extraction from BC database. The rules are generated in this system which are most transparent to verify the Major Risk factors of BC. The proposed TNESBC algorithm is comparable the RxREN algorithm. The experimental analysis, it is noticed that TNESBC algorithm performs better than RxREN. The comparison is formed using 10-fold Cross-Validation(CV) accuracy, average recall rate, average number of antecedents per rules and average false positive rate.

Representing the knowledge learned by neural networks in the form of interpretable rules is a prudent technique to justify the decisions made by neural networks. Heretofore many algorithms exist to extract symbolic rules from neural networks, but among them, a few extract rules from deep neural networks trained using deep learning techniques. So, this paper proposes an algorithm to extract rules from a multi-hidden layer neural network, pre-trained using deep belief network and fine-tuned using back propagation. The algorithm analyzes each node of a layer and extracts knowledge from each layer separately. The process of knowledge extraction from the first hidden layer is different from the other layers. Consecutively, the algorithm combines all the knowledge extracted and refines them to construct a final ruleset consisting of symbolic rules. The algorithm further subdivides the subspace of a rule in the ruleset if it satisfies certain conditions. Results show that the algorithm extracted rules with higher accuracy compared to some existing rule extraction algorithms. Other than accuracy, the efficacy of the extracted rules is also validated with fidelity and various other performance measures.

Nowadays credit risk evaluation is very crucial in financial domain. Whenever it is processed by an individual, it becomes controversial as the assessment may be prone to human error. Recently, to overcome this issue, some automated systems have been developed for credit risk evaluation. Most of the developed systems focused on the credit decision only and neglected the transparency of the systems; however, many cases require transparency of the credit decision to benefit financial organization as well as the potential customers. Therefore, this paper proposes an expert system named Transparent Neural based Expert System for Credit Risk (TNESCR) evaluation which uses a white box neural model Rule Extraction from Neural Network using Classified and Misclassified data (RxNCM) to generate rules from financial data. The generated rules are so transparent to justify the explanations for why applications are granted/rejected with a significant predictive accuracy. The proposed TNESCR is validated using 10 fold cross validation with 3 credit risk datasets. The experimental results show the proposed TNESCR can perform significantly with great transparency and accuracy.

The increase in the rate of technological evolution is resulting in a reduction in the cost of various storage devices and as a consequence, enormous amounts of data are deposited from heterogeneous sources in raw forms. Therefore, some efficient data mining techniques are required that can process those data and retrieve useful information from them. Recently, machine learning algorithms are becoming very popular for doing various data mining tasks. Neural network is one of them, which has fascinated a lot of researchers due to its efficacy and fruitfulness in doing many tasks specially classification. But the main problem with neural network is its nature of black box, i.e., explaining the decision generated by a neural network is a daunting task. As a solution to this problem, rule extraction technique has been proposed which expresses the knowledge hidden in a learned network in the guise of understandable classification rules. The rule extraction is a very deep rooted technique and a very rich literature exists on this topic. However, a very less number of papers exist which mainly focused on surveying the existing literature. So, this work aims to provide a survey on the existing literature, and to shed light on some of the areas which needs to be focused to enrich the literature. At the same time the paper also tries to create a scope for the existing and the novice researchers to do research in this field.

Classification is one of the important tasks in data mining. Over the last few decades, neural networks have proved to be very effective in solving this task. This paper presents a novel algorithm, called Neural Network Boosting (NNBOOST) to improve the classification performance of neural networks using ensembling technique. The ensembling technique proposed here is based on the boosting concept to ensemble classifiers. A boosting algorithm sequentially learns a new model by reassigning pattern weights. NNBOOST algorithm adapts a new technique to assign weights to patterns based on the Euclidean distances of patterns from the centroids of respective classes. It assigns more weight to patterns closest to the centroid. The pattern weights are updated subsequently after building or learning a neural network model, based on the updated centroids to build the next neural network model. Weighted Majority voting scheme is used to combine all the learned neural network models. Weight or confidence of a learned neural network model is calculated as the ratio of correctly classified to misclassified training patterns. The algorithm is validated with six real life data sets taken from UCI repository. Results show the effectiveness of the algorithm in ensembling neural networks.

Diabetic retinopathy (DR) or diabetic eye disease is a medical condition in which damage occurs to the retina due to diabetes mellitus causing blindness. Early detection of DR helps in preventing the onset of blindness in diabetic patients. Numerous intelligent models have been proposed for early detection of DR. However, models incorporating support vector machine (SVM) for early detection of DR have been very rare. Therefore, this paper proposes a model called intelligent system for diabetic retinopathy (ISDR) for early detection of DR using SVM. The fundus images captured with the help of a digital fundus camera (DFC) are used as inputs to the proposed model. The fundus images are initially enhanced and then segmented to extract two most prominent features for the early detection of DR, namely foveal avascular zone (FAZ) and microaneurysms (MA). The extracted features are used as inputs to SVM for classification of the fundus images to classify them as 0—no DR, 1—mild DR, 2—moderate DR, 3—non-proliferative DR (NPDR) and 4—proliferative DR (PDR). The model is validated by the Kaggle dataset. It is observed from the experimental results that early detection of DR is feasible using ISDR.

Neural network is one of the best tools for data mining tasks due to its high accuracy. However, one of the drawbacks of neural network is its black box nature. This limitation makes neural network useless for many applications which require transparency in their decision-making process. Many algorithms have been proposed to overcome this drawback by extracting transparent rules from neural network, but still researchers are in search for algorithms that can generate more accurate and simple rules. Therefore, this paper proposes a rule extraction algorithm named Eclectic Rule Extraction from Neural Network Recursively (ERENNR), with the aim to generate simple and accurate rules. ERENNR algorithm extracts symbolic classification rules from a single-layer feed-forward neural network. The novelty of this algorithm lies in its procedure of analyzing the nodes of the network. It analyzes a hidden node based on data ranges of input attributes with respect to its output and analyzes an output node using logical combination of the outputs of hidden nodes with respect to output class. And finally it generates a rule set by proceeding in a backward direction starting from the output layer. For each rule in the set, it repeats the whole process of rule extraction if the rule satisfies certain criteria. The algorithm is validated with eleven benchmark datasets. Experimental results show that the generated rules are simple and accurate.

This paper aims at presenting a survey of the ongoing research on use of various social media platforms to manage various phases of a crisis situation. A thorough research has been carried out on the existing works and their limitations and scope of future have been highlighted to understand the progress made so far. The aim is to identify the areas that still require attention and to help build a system which aims at higher accuracy in information filtering and is robust to meet the real-time information demands to counter the crisis using social media platforms. The research has been categorized into two parts: Analysis of situational awareness information propagated during a crisis situation and second, information sharing among various communities and understanding community’s role during disasters.

Nowadays, people face various health related problems due to the busy Iifestyle and hectic work pressure. Under this circumstance, if it is possible to have a system that identifies the major causes of health related problems then people can follow some precautionary measures to prevent those causes and can have disease free life. This paper proposes an algorithm, called Rule Based Major Feature Identification (RBMFI) which identifies major feature of medical problems by pruning p-rules generated by decision tree. RBMFI algorithm comprises of three phases: Rule generation, Rule pruning and Attribute pruning. Rule generation generates the rules using decision tree, rule pruning removes the rules that do not affect the overall classification accuracy and attribute pruning keeps only the major feature of a decision. Finally, the major feature is treated as major cause of a decision (disease). The proposed method is validated with three medical datasets taken from UCI repository. It is observed from the experimental results that the proposed RBMFI can identify the major feature with good accuracy.

This article describes how the enormous size of data in IoT needs efficient data mining model for information extraction, classification and mining hidden patterns from data. CBR is a learning, mining and problem-solving approach which solves a problem by relating past similar solved problems. One issue with CBR is feature weight to measure the similarity among cases to mine similar past cases. NN's pruning is a popular method, which extracts feature weights from a trained neural network without losing much generality of the training set by using four mechanisms: sensitivity, activity, saliency and relevance. However, training NN with imbalanced data leads the classifier to get biased towards the majority class. Therefore, this article proposes a hybrid CBR model with RUS and cost sensitive back propagation neural network in IoT environment to deal with the feature weighting problem in imbalance data. The proposed model is validated with six real-life datasets. The experimental results show that the proposed model is better than other feature weighting methods.

This paper proposes an algorithm named Reverse Engineering Recursive Rule Extraction (RE-Re-RX) for symbolic rule extraction from neural network with mixed attributes. RE-Re-RX algorithm is an extension of the existing Recursive Rule Extraction (Re-RX) algorithm. Re-RX algorithm generates disjoint rules for continuous and discrete attributes. The algorithm first generates rules for discrete attributes. A rule for discrete attributes is further refined recursively if it does not produce satisfactory result. A rule is refined by generating rules with the discrete attributes (if present) that are not covered by the rule or else the process is terminated by generating rules with the continuous attributes (if present). The novelty of the proposed RE-Re-RX algorithm lies in generating rules for continuous attributes. Re-RX generates linear hyper plane for continuous attributes which may not be able to deal with the non-linearity present in data. To overcome this limitation RE-Re-RX algorithm generates simple rules for continuous attributes in the form of input data ranges and target. RE-Re-RX uses the concept of Rule Extraction by Reverse Engineering the NN (RxREN) algorithm in slightly different way to generate rules. RxREN only uses misclassified patterns, whereas RE-Re-RX uses both classified and misclassified patterns of each continuous attribute to calculate input data ranges for constructing rules. The proposed algorithm is validated with six benchmark datasets. The experimental results clearly show the superiority of the proposed algorithm to Re-RX.

Classification is one of the important tasks of data mining and neural network is one of the best known tools for doing this task. Despite of producing high classification accuracy, the black box nature of neural network makes it useless for many applications which require transparency in its decision-making process. This drawback is overcome by extracting rules from neural network. Rule extraction makes neural network an alternative to other machine learning methods for handling classification problems by deriving an explanation of how each decision is made. Till now, many algorithms on rule extraction have been proposed but still research on this area is going on to find out more accurate and understandable rules. The proposed algorithm extracts rules from trained neural network for datasets with mixed mode attributes using pedagogical approach. The proposed algorithm uses classified and misclassified patterns to find out the data ranges of significant attributes in respective classes. The experimental results clearly show that the proposed algorithm produces accurate and understandable rules compared to existing algorithms.

Case-based reasoning (CBR) is an artificial intelligent approach to learning and problem-solving, which solves a target problem by relating past similar solved problems. But it faces the challenge of weights assignment to features to measure similarity between cases. There are many methods to overcome this feature weighting problem of CBR. However, neural network’s pruning is one of the powerful and useful methods to overcome this feature weighting problem, which extracts feature weights from trained neural network without losing the generality of training set by four popular mechanisms: sensitivity, activity, saliency and relevance. It is habitually assumed that the training sets used for learning are balanced. However, this hypothesis is not always true in real-world applications, and hence, the tendency is to yield classification models that are biased toward the overrepresented class. Therefore, a hybrid CBR system is proposed in this paper to overcome this problem, which adopts a cost-sensitive back-propagation neural network (BPNN) in network pruning to find feature weights. These weights are used in CBR. A single cost parameter is used by the cost-sensitive BPNN to distinguish the importance of class errors. A balanced decision boundary is generated by the cost parameter using prior information. Thus, the class imbalance problem of network pruning is overcome to improve the accuracy of the hybrid CBR. From the empirical results, it is observed that the performance of the proposed hybrid CBR system is better than the hybrid CBR by standard neural network. The performance of the proposed hybrid system is validated with seven datasets.

Data Mining is a powerful technology to help organization to concentrate on most important data by extracting useful information from large database. One of the most commonly used techniques in data mining is Artificial Neural Network due to its high performance in many application domains. Despite many advantages of Artificial Neural Network, one of its main drawbacks is its inherent black box nature which is the main problem of using Artificial Neural Network in data mining. Therefore, this paper proposes a rule extraction algorithm from neural network using classified and misclassified data to convert the black box nature of Artificial Neural Network into a white box. The proposed algorithm is a modification of the existing algorithm, Rule Extraction by Reverse Engineering (RxREN). The proposed algorithm extracts rules from trained neural network for datasets with mixed mode attributes using pedagogical approach. The proposed algorithm uses both classified as well as misclassified data to find out the data ranges of significant attributes in respective classes, which is the innovation of the proposed algorithm. The experimental results clearly show that the performance of the proposed algorithm is superior to existing algorithms.