List of Articles Machine Learning

• Open Access Article
  • Abstract Page
  • Full-Text
  
  1 - Information Bottleneck and its Applications in Deep Learning
  Hassan Hafez Kolahi Shohreh Kasaei
  10.7508/jist.2018.03.001
  Information Theory (IT) has been used in Machine Learning (ML) from early days of this field. In the last decade, advances in Deep Neural Networks (DNNs) have led to surprising improvements in many applications of ML. The result has been a paradigm shift in the communit More

  Information Theory (IT) has been used in Machine Learning (ML) from early days of this field. In the last decade, advances in Deep Neural Networks (DNNs) have led to surprising improvements in many applications of ML. The result has been a paradigm shift in the community toward revisiting previous ideas and applications in this new framework. Ideas from IT are no exception. One of the ideas which is being revisited by many researchers in this new era, is Information Bottleneck (IB); a formulation of information extraction based on IT. The IB is promising in both analyzing and improving DNNs. The goal of this survey is to review the IB concept and demonstrate its applications in deep learning. The information theoretic nature of IB, makes it also a good candidate in showing the more general concept of how IT can be used in ML. Two important concepts are highlighted in this narrative on the subject, i) the concise and universal view that IT provides on seemingly unrelated methods of ML, demonstrated by explaining how IB relates to minimal sufficient statistics, stochastic gradient descent, and variational auto-encoders, and ii) the common technical mistakes and problems caused by applying ideas from IT, which is discussed by a careful study of some recent methods suffering from them. Manuscript profile
• Open Access Article
  • Abstract Page
  • Full-Text
  
  2 - Density Measure in Context Clustering for Distributional Semantics of Word Sense Induction
  Masood Ghayoomi
  10.7508/jist.2020.01.002
  Word Sense Induction (WSI) aims at inducing word senses from data without using a prior knowledge. Utilizing no labeled data motivated researchers to use clustering techniques for this task. There exist two types of clustering algorithm: parametric or non-parametric. Al More

  Word Sense Induction (WSI) aims at inducing word senses from data without using a prior knowledge. Utilizing no labeled data motivated researchers to use clustering techniques for this task. There exist two types of clustering algorithm: parametric or non-parametric. Although non-parametric clustering algorithms are more suitable for inducing word senses, their shortcomings make them useless. Meanwhile, parametric clustering algorithms show competitive results, but they suffer from a major problem that is requiring to set a predefined fixed number of clusters in advance. Word Sense Induction (WSI) aims at inducing word senses from data without using a prior knowledge. Utilizing no labeled data motivated researchers to use clustering techniques for this task. There exist two types of clustering algorithm: parametric or non-parametric. Although non-parametric clustering algorithms are more suitable for inducing word senses, their shortcomings make them useless. Meanwhile, parametric clustering algorithms show competitive results, but they suffer from a major problem that is requiring to set a predefined fixed number of clusters in advance. The main contribution of this paper is to show that utilizing the silhouette score normally used as an internal evaluation metric to measure the clusters’ density in a parametric clustering algorithm, such as K-means, in the WSI task captures words’ senses better than the state-of-the-art models. To this end, word embedding approach is utilized to represent words’ contextual information as vectors. To capture the context in the vectors, we propose two modes of experiments: either using the whole sentence, or limited number of surrounding words in the local context of the target word to build the vectors. The experimental results based on V-measure evaluation metric show that the two modes of our proposed model beat the state-of-the-art models by 4.48% and 5.39% improvement. Moreover, the average number of clusters and the maximum number of clusters in the outputs of our proposed models are relatively equal to the gold data Manuscript profile
• Open Access Article
  • Abstract Page
  • Full-Text
  
  3 - A Study of Fraud Types, Challenges and Detection Approaches in Telecommunication
  Kasra Babaei ZhiYuan Chen Tomas Maul
  10.7508/jist.2019.04.002
  Fraudulent activities have been rising globally resulting companies losing billions of dollars that can cause severe financial damages. Various approaches have been proposed by researchers in different applications. Studying these approaches can help us obtain a better More

  Fraudulent activities have been rising globally resulting companies losing billions of dollars that can cause severe financial damages. Various approaches have been proposed by researchers in different applications. Studying these approaches can help us obtain a better understanding of the problem. The aim of this paper is to investigate different aspects of fraud prevention and detection in telecommunication. This study presents a review of different fraud categories in telecommunication, the challenges that hinder the detection process, and some proposed solutions to overcome them. Also, the performance of some of the state-of-the-art approaches is reported followed by our guideline and recommendation in choosing the best metrics. Manuscript profile
• Open Access Article
  • Abstract Page
  • Full-Text
  
  4 - An Effective Method of Feature Selection in Persian Text for Improving the Accuracy of Detecting Request in Persian Messages on Telegram
  zahra khalifeh zadeh Mohammad Ali Zare Chahooki
  10.29252/jist.8.32.249
  20.1001.1.23221437.2020.8.32.6.2
  In recent years, data received from social media has increased exponentially. They have become valuable sources of information for many analysts and businesses to expand their business. Automatic document classification is an essential step in extracting knowledge from More

  In recent years, data received from social media has increased exponentially. They have become valuable sources of information for many analysts and businesses to expand their business. Automatic document classification is an essential step in extracting knowledge from these sources of information. In automatic text classification, words are assessed as a set of features. Selecting useful features from each text reduces the size of the feature vector and improves classification performance. Many algorithms have been applied for the automatic classification of text. Although all the methods proposed for other languages are applicable and comparable, studies on classification and feature selection in the Persian text have not been sufficiently carried out. The present research is conducted in Persian, and the introduction of a Persian dataset is a part of its innovation. In the present article, an innovative approach is presented to improve the performance of Persian text classification. The authors extracted 85,000 Persian messages from the Idekav-system, which is a Telegram search engine. The new idea presented in this paper to process and classify this textual data is on the basis of the feature vector expansion by adding some selective features using the most extensively used feature selection methods based on Local and Global filters. The new feature vector is then filtered by applying the secondary feature selection. The secondary feature selection phase selects more appropriate features among those added from the first step to enhance the effect of applying wrapper methods on classification performance. In the third step, the combined filter-based methods and the combination of the results of different learning algorithms have been used to achieve higher accuracy. At the end of the three selection stages, a method was proposed that increased accuracy up to 0.945 and reduced training time and calculations in the Persian dataset. Manuscript profile
• Open Access Article
  • Abstract Page
  • Full-Text
  
  5 - Evaluation of Pattern Recognition Techniques in Response to Cardiac Resynchronization Therapy (CRT)
  Mohammad Nejadeh Peyman Bayat Jalal Kheirkhah Hassan Moladoust
  10.29252/jist.8.31.197
  20.1001.1.23221437.2020.8.31.2.6
  Cardiac resynchronization therapy (CRT) improves cardiac function in patients with heart failure (HF), and the result of this treatment is decrease in death rate and improving quality of life for patients. This research is aimed at predicting CRT response for the progno More

  Cardiac resynchronization therapy (CRT) improves cardiac function in patients with heart failure (HF), and the result of this treatment is decrease in death rate and improving quality of life for patients. This research is aimed at predicting CRT response for the prognosis of patients with heart failure under CRT. According to international instructions, in the case of approval of QRS prolongation and decrease in ejection fraction (EF), the patient is recognized as a candidate of implanting recognition device. However, regarding many intervening and effective factors, decision making can be done based on more variables. Computer-based decision-making systems especially machine learning (ML) are considered as a promising method regarding their significant background in medical prediction. Collective intelligence approaches such as particles swarm optimization (PSO) algorithm are used for determining the priorities of medical decision-making variables. This investigation was done on 209 patients and the data was collected over 12 months. In HESHMAT CRT center, 17.7% of patients did not respond to treatment. Recognizing the dominant parameters through combining machine recognition and physician’s viewpoint, and introducing back-propagation of error neural network algorithm in order to decrease classification error are the most important achievements of this research. In this research, an analytical set of individual, clinical, and laboratory variables, echocardiography, and electrocardiography (ECG) are proposed with patients’ response to CRT. Prediction of the response after CRT becomes possible by the support of a set of tools, algorithms, and variables. Manuscript profile
• Open Access Article
  • Abstract Page
  • Full-Text
  
  6 - Predicting Student Performance for Early Intervention using Classification Algorithms in Machine Learning
  Kalaivani K Ulagapriya K Saritha A Ashutosh  Kumar
  10.52547/jist.9.36.226
  20.1001.1.23221437.2021.9.36.1.2
  Predicting Student’s Performance System is to find students who may require early intervention before they fail to graduate. It is generally meant for the teaching faculty members to analyze Student's Performance and Results. It stores Student Details in a database and More

  Predicting Student’s Performance System is to find students who may require early intervention before they fail to graduate. It is generally meant for the teaching faculty members to analyze Student's Performance and Results. It stores Student Details in a database and uses Machine Learning Model using i. Python Data Analysis tools like Pandas and ii. Data Visualization tools like Seaborn to analyze the overall Performance of the Class. The proposed system suggests student performance prediction through Machine Learning Algorithms and Data Mining Techniques. The Data Mining technique used here is classification, which classifies the students based on student’s attributes. The Front end of the application is made using React JS Library with Data Visualization Charts and connected to a backend Database where all student’s records are stored in MongoDB and the Machine Learning model is trained and deployed through Flask. In this process, the machine learning algorithm is trained using a dataset to create a model and predict the output on the basis of that model. Three different types of data used in Machine Learning are continuous, categorical and binary. In this study, a brief description and comparative analysis of various classification techniques is done using student performance dataset. The six different machine learning Classification algorithms, which have been compared, are Logistic Regression, Decision Tree, K-Nearest Neighbor, Naïve Bayes, Support Vector Machine and Random Forest. The results of Naïve Bayes classifier are comparatively higher than other techniques in terms of metrics such as precision, recall and F1 score. The values of precision, recall and F1 score are 0.93, 0.92 and 0.92 respectively. Manuscript profile
• Open Access Article
  • Abstract Page
  • Full-Text
  
  7 - A Hybrid Machine Learning Approach for Sentiment Analysis of Beauty Products Reviews
  Kanika Jindal Rajni Aron
  10.52547/jist.15586.10.37.1
  20.1001.1.23221437.2022.10.37.1.1
  Nowadays, social media platforms have become a mirror that imitates opinions and feelings about any specific product or event. These product reviews are capable of enhancing communication among entrepreneurs and their customers. These reviews need to be extracted and an More

  Nowadays, social media platforms have become a mirror that imitates opinions and feelings about any specific product or event. These product reviews are capable of enhancing communication among entrepreneurs and their customers. These reviews need to be extracted and analyzed to predict the sentiment polarity, i.e., whether the review is positive or negative. This paper aims to predict the human sentiments expressed for beauty product reviews extracted from Amazon and improve the classification accuracy. The three phases instigated in our work are data pre-processing, feature extraction using the Bag-of-Words (BoW) method, and sentiment classification using Machine Learning (ML) techniques. A Global Optimization-based Neural Network (GONN) is proposed for the sentimental classification. Then an empirical study is conducted to analyze the performance of the proposed GONN and compare it with the other machine learning algorithms, such as Random Forest (RF), Naive Bayes (NB), and Support Vector Machine (SVM). We dig further to cross-validate these techniques by ten folds to evaluate the most accurate classifier. These models have also been investigated on the Precision-Recall (PR) curve to assess and test the best technique. Experimental results demonstrate that the proposed method is the most appropriate method to predict the classification accuracy for our defined dataset. Specifically, we exhibit that our work is adept at training the textual sentiment classifiers better, thereby enhancing the accuracy of sentiment prediction. Manuscript profile
• Open Access Article
  • Abstract Page
  • Full-Text
  
  8 - The Development of a Hybrid Error Feedback Model for Sales Forecasting
  Mehdi Farrokhbakht Foumani Sajad Moazami Goudarzi
  10.52547/jist.9.34.131
  20.1001.1.23221437.2021.9.34.7.4
  Sales forecasting is one of the significant issues in the industrial and service sector which can lead to facilitated management decisions and reduce the lost values in case of being dealt with properly. Also sales forecasting is one of the complicated problems in analy More

  Sales forecasting is one of the significant issues in the industrial and service sector which can lead to facilitated management decisions and reduce the lost values in case of being dealt with properly. Also sales forecasting is one of the complicated problems in analyzing time series and data mining due to the number of intervening parameters. Various models were presented on this issue and each one found acceptable results. However, developing the methods in this study is still considered by researchers. In this regard, the present study provided a hybrid model with error feedback for sales forecasting. In this study, forecasting was conducted using a supervised learning method. Then, the remaining values (model error) were specified and the error values were forecasted using another learning method. Finally, two trained models were combined together and consecutively used for sales forecasting. In other words, first the forecasting was conducted and then the error rate was determined by the second model. The total forecasting and model error indicated the final forecasting. The computational results obtained from numerical experiments indicated the superiority of the proposed hybrid method performance over the common models in the available literature and reduced the indicators related to forecasting error. Manuscript profile
• Open Access Article
  • Abstract Page
  • Full-Text
  
  9 - Word Sense Induction in Persian and English: A Comparative Study
  Masood Ghayoomi
  10.52547/jist.9.36.263
  20.1001.1.23221437.2021.9.36.3.4
  Words in the natural language have forms and meanings, and there might not always be a one-to-one match between them. This property of the language causes words to have more than one meaning; as a result, a text processing system faces challenges to determine the precis More

  Words in the natural language have forms and meanings, and there might not always be a one-to-one match between them. This property of the language causes words to have more than one meaning; as a result, a text processing system faces challenges to determine the precise meaning of the target word in a sentence. Using lexical resources or lexical databases, such as WordNet, might be a help, but due to their manual development, they become outdated by passage of time and language change. Moreover, the lexical resources might be domain dependent which are unusable for open domain natural language processing tasks. These drawbacks are a strong motivation to use unsupervised machine learning approaches to induce word senses from the natural data. To reach the goal, the clustering approach can be utilized such that each cluster resembles a sense. In this paper, we study the performance of a word sense induction model by using three variables: a) the target language: in our experiments, we run the induction process on Persian and English; b) the type of the clustering algorithm: both parametric clustering algorithms, including hierarchical and partitioning, and non-parametric clustering algorithms, including probabilistic and density-based, are utilized to induce senses; c) the context of the target words to capture the information in vectors created for clustering: for the input of the clustering algorithms, the vectors are created either based on the whole sentence in which the target word is located; or based on the limited surrounding words of the target word. We evaluate the clustering performance externally. Moreover, we introduce a normalized, joint evaluation metric to compare the models. The experimental results for both Persian and English test data showed that the window-based partitioningK-means algorithm obtained the best performance. Manuscript profile
• Open Access Article
  • Abstract Page
  • Full-Text
  
  10 - Deep Learning-based Educational User Profile and User Rating Recommendation System for E-Learning
  Pradnya Vaibhav  Kulkarni Sunil Rai Rajneeshkaur Sachdeo Rohini Kale
  10.61186/jist.27448.11.43.185
  20.1001.1.23221437.2023.11.43.3.2
  In the current era of online learning, the recommendation system for the eLearning process is quite important. Since the COVID-19 pandemic, eLearning has undergone a complete transformation. Existing eLearning Recommendation Systems worked on collaborative filtering or More

  In the current era of online learning, the recommendation system for the eLearning process is quite important. Since the COVID-19 pandemic, eLearning has undergone a complete transformation. Existing eLearning Recommendation Systems worked on collaborative filtering or content-based filtering based on historical data, students’ previous grade, results, or user profiles. The eLearning system selected courses based on these parameters in a generalized manner rather than on a personalized basis. Personalized recommendations, information relevancy, choosing the proper course, and recommendation accuracy are some of the issues in eLearning recommendation systems. In this paper, existing conventional eLearning and course recommendation systems are studied in detail and compared with the proposed approach. We have used, the dataset of User Profile and User Rating for a recommendation of the course. K Nearest Neighbor, Support Vector Machine, Decision Tree, Random Forest, Nave Bayes, Linear Regression, Linear Discriminant Analysis, and Neural Network were among the Machine Learning techniques explored and deployed. The accuracy achieved for all these algorithms ranges from 0.81 to 0.97. The proposed algorithm uses a hybrid approach by combining collaborative filtering and deep learning. We have improved accuracy to 0.98 which indicate that the proposed model can provide personalized and accurate eLearning recommendation for the individual user. Manuscript profile
• Open Access Article
  • Abstract Page
  • Full-Text
  
  11 - An Autoencoder based Emotional Stress State Detection Approach by using Electroencephalography Signals
  Jia Uddin
  10.52547/jist.32267.11.41.24
  20.1001.1.23221437.2023.11.41.6.1
  Identifying hazards from human error is critical for industrial safety since dangerous and reckless industrial worker actions, as well as a lack of measures, are directly accountable for human-caused problems. Lack of sleep, poor nutrition, physical deformities, and wea More

  Identifying hazards from human error is critical for industrial safety since dangerous and reckless industrial worker actions, as well as a lack of measures, are directly accountable for human-caused problems. Lack of sleep, poor nutrition, physical deformities, and weariness are some of the key factors that contribute to these risky and reckless behaviors that might put a person in a perilous scenario. This scenario causes discomfort, worry, despair, cardiovascular disease, a rapid heart rate, and a slew of other undesirable outcomes. As a result, it would be advantageous to recognize people's mental states in the future in order to provide better care for them. Researchers have been studying electroencephalogram (EEG) signals to determine a person's stress level at work in recent years. A full feature analysis from domains is necessary to develop a successful machine learning model using electroencephalogram (EEG) inputs. By analyzing EEG data, a time-frequency based hybrid bag of features is designed in this research to determine human stress dependent on their sex. This collection of characteristics includes features from two types of assessments: time-domain statistical analysis and frequency-domain wavelet-based feature assessment. The suggested two layered autoencoder based neural networks (AENN) are then used to identify the stress level using a hybrid bag of features. The experiment uses the DEAP dataset, which is freely available. The proposed method has a male accuracy of 77.09% and a female accuracy of 80.93%. Manuscript profile
• Open Access Article
  • Abstract Page
  • Full-Text
  
  12 - Comparative Study of 5G Signal Attenuation Estimation Models
  Md Anoarul Islam Manabendra Maiti Judhajit Sanyal Quazi Md Alfred
  10.52547/jist.32544.11.42.84
  20.1001.1.23221437.2023.11.42.4.1
  Wireless networks functioning on 4G and 5G technology offer a plethora of options to users in terms of connectivity and multimedia content. However, such networks are prone to severe signal attenuation and noise in a number of scenarios. Significant research in recent y More

  Wireless networks functioning on 4G and 5G technology offer a plethora of options to users in terms of connectivity and multimedia content. However, such networks are prone to severe signal attenuation and noise in a number of scenarios. Significant research in recent years has consequently focused on establishment of robust and accurate attenuation models to estimate channel noise and subsequent signal loss. The identified challenge therefore is to identify or develop accurate computationally inexpensive models implementable on available hardware for generation of estimates with low error and validate the solutions experimentally. The present work surveys some of the most relevant recent work in this domain, with added emphasis on rain attenuation models and machine learning based approaches, and offers a perspective on the establishment of a suitable dynamic signal attenuation model for high-speed wireless communication in outdoor as well as indoor environments, presenting the performance evaluation of an autoregression-based machine learning model. Multiple versions of the model are compared on the basis of root mean square error (RMSE) for different orders of regression polynomials to find the best-fit solution. The accuracy of the technique proposed in the paper is then compared in terms of RMSE to corresponding moderate and high complexity machine learning techniques implementing adaptive spline regression and artificial neural networks respectively. The proposed method is found to be quite accurate with low complexity, allowing the method to be practically applicable in multiple scenarios. Manuscript profile
• Open Access Article
  • Abstract Page
  • Full-Text
  
  13 - Breast Cancer Classification Approaches - A Comparative Analysis
  Mohan Kumar Sunil Kumar Khatri Masoud Mohammadian
  10.52547/jist.33395.11.41.1
  20.1001.1.23221437.2023.11.41.8.3
  Cancer of the breast is a difficult disease to treat since it weakens the patient's immune system. Particular interest has lately been shown in the identification of particular immune signals for a variety of malignancies in this regard. In recent years, several methods More

  Cancer of the breast is a difficult disease to treat since it weakens the patient's immune system. Particular interest has lately been shown in the identification of particular immune signals for a variety of malignancies in this regard. In recent years, several methods for predicting cancer based on proteomic datasets and peptides have been published. The cells turns into cancerous cells because of various reasons and get spread very quickly while detrimental to normal cells. In this regard, identifying specific immunity signs for a range of cancers has recently gained a lot of interest. Accurately categorizing and compartmentalizing the breast cancer subtype is a vital job. Computerized systems built on artificial intelligence can substantially save time and reduce inaccuracy. Several strategies for predicting cancer utilizing proteomic datasets and peptides have been reported in the literature in recent years.It is critical to classify and categorize breast cancer treatments correctly. It's possible to save time while simultaneously minimizing the likelihood of mistakes using machine learning and artificial intelligence approaches. Using the Wisconsin Breast Cancer Diagnostic dataset, this study evaluates the performance of various classification methods, including SVC, ETC, KNN, LR, and RF (random forest). Breast cancer can be detected and diagnosed using a variety of measurements of data (which are discussed in detail in the article) (WBCD). The goal is to determine how well each algorithm performs in terms of precision, recall, and accuracy. The variation of each classification threshold has been tested on various algorithms and SVM turned out to be very promising. Manuscript profile
• Open Access Article
  • Abstract Page
  • Full-Text
  
  14 - Implementation of Machine Learning Algorithms for Customer Churn Prediction
  Manal Loukili Fayçal Messaoudi Raouya El Youbi
  10.61186/jist.34208.11.43.196
  20.1001.1.23221437.2023.11.43.5.4
  Churn prediction is one of the most critical issues in the telecommunications industry. The possibilities of predicting churn have increased considerably due to the remarkable progress made in the field of machine learning and artificial intelligence. In this context, w More

  Churn prediction is one of the most critical issues in the telecommunications industry. The possibilities of predicting churn have increased considerably due to the remarkable progress made in the field of machine learning and artificial intelligence. In this context, we propose the following process which consists of six stages. The first phase consists of data pre-processing, followed by feature analysis. In the third phase, the selection of features. Then the data was divided into two parts: the training set and the test set. In the prediction process, the most popular predictive models were adopted, namely random forest, k-nearest neighbor, and support vector machine. In addition, we used cross-validation on the training set for hyperparameter tuning and to avoid model overfitting. Then, the results obtained on the test set were evaluated using the confusion matrix and the AUC curve. Finally, we found that the models used gave high accuracy values (over 79%). The highest AUC score, 84%, is achieved by the SVM and bagging classifiers as an ensemble method which surpasses them. Manuscript profile
• Open Access Article
  • Abstract Page
  • Full-Text
  
  15 - An Analysis of Covid-19 Pandemic Outbreak on Economy using Neural Network and Random Forest
  Md. Nahid  Hasan Tanvir  Ahmed Md.  Ashik Md. Jahid  Hasan Tahaziba  Azmin Jia Uddin
  10.52547/jist.34246.11.42.163
  20.1001.1.23221437.2023.11.42.7.4
  The pandemic disease outbreaks are causing a significant financial crisis affecting the worldwide economy. Machine learning techniques are urgently required to detect, predict and analyze the economy for early economic planning and growth. Consequently, in this paper, w More

  The pandemic disease outbreaks are causing a significant financial crisis affecting the worldwide economy. Machine learning techniques are urgently required to detect, predict and analyze the economy for early economic planning and growth. Consequently, in this paper, we use machine learning classifiers and regressors to construct an early warning model to tackle economic recession due to the cause of covid-19 pandemic outbreak. A publicly available database created by the National Bureau of Economic Research (NBER) is used to validate the model, which contains information about national revenue, employment rate, and workers' earnings of the USA over 239 days (1 January 2020 to 12 May 2020). Different techniques such as missing value imputation, k-fold cross validation have been used to pre-process the dataset. Machine learning classifiers- Multi-layer Perceptron- Neural Network (MLP-NN) and Random Forest (RF) have been used to predict recession. Additionally, machine learning regressors-Long Short-Term Memory (LSTM) and Random Forest (RF) have been used to detect how much recession a country is facing as a result of positive test cases of covid-19 pandemic. Experimental results demonstrate that the MLP-NN and RF classifiers have exhibited average 88.33% and 85% of recession (where 95%, 81%, 89% and 85%, 81%, 89% for revenue, employment rate and workers earnings, respectively) and average 90.67% and 93.67% of prediction accuracy for LSTM and RF regressors (where 92%, 90%, 90%, and 95%, 93%, 93% respectively). Manuscript profile
• Open Access Article
  • Abstract Page
  • Full-Text
  
  16 - Application of Machine Learning in the Telecommunications Industry: Partial Churn Prediction by using a Hybrid Feature Selection Approach
  Fatemeh Mozaffari Iman Raeesi Vanani Payam Mahmoudian Babak Sohrabi
  10.61186/jist.38419.11.44.331
  The telecommunications industry is one of the most competitive industries in the world. Because of the high cost of customer acquisition and the adverse effects of customer churn on the company's performance, customer retention becomes an inseparable part of strategic d More

  The telecommunications industry is one of the most competitive industries in the world. Because of the high cost of customer acquisition and the adverse effects of customer churn on the company's performance, customer retention becomes an inseparable part of strategic decision-making and one of the main objectives of customer relationship management. Although customer churn prediction models are widely studied in various domains, several challenges remain in designing and implementing an effective model. This paper addresses the customer churn prediction problem with a practical approach. The experimental analysis was conducted on the customers' data gathered from available sources at a telecom company in Iran. First, partial churn was defined in a new way that exploits the status of customers based on criteria that can be measured easily in the telecommunications industry. This definition is also based on data mining techniques that can find the degree of similarity between assorted customers with active ones or churners. Moreover, a hybrid feature selection approach was proposed in which various feature selection methods, along with the crowd's wisdom, were applied. It was found that the wisdom of the crowd can be used as a useful feature selection method. Finally, a predictive model was developed using advanced machine learning algorithms such as bagging, boosting, stacking, and deep learning. The partial customer churn was predicted with more than 88% accuracy by the Gradient Boosting Machine algorithm by using 5-fold cross-validation. Comparative results indicate that the proposed model performs efficiently compared to the ones applied in the previous studies. Manuscript profile