Repeated Incremental Pruning To Produce Error Reduction Research Articles

Sampling-based novel heterogeneous multi-layer stacking ensemble method for telecom customer churn prediction

In recent times, customer churn has become one of the most significant issues in business-oriented sectors with telecommunication being no exception. Maintaining current customers is particularly valuable due to the high degree of rivalry among telecommunication companies and the costs of acquiring new ones. The early prediction of churned customers may help telecommunication companies to identify the causes of churn and design industrial tactics to address or mitigate the churn problem. Controlling customer churn by developing efficient and reliable customer churn prediction (CCP) solutions is essential to achieving this objective. Findings from existing CCP studies have shown that numerous methods, such as rule-based and machine-learning (ML) mechanisms, have been devised to solve the CCP problem. Nonetheless, the problems of adaptability and the resilience of rule-based CCP solutions are its major weaknesses, and the skewed pattern of churn datasets (class imbalance) is detrimental to the prediction performances of conventional ML models in CCP. Hence, this research developed a robust heterogeneous multi-layer stacking ensemble method (HMSE) for effective CCP. Specifically, in the HMSE method, the prediction prowess of five ML classifiers (Random Forest (RF), Bayesian network (BN), Support Vector Machine (SVM), K-Nearest Neighbour (KNN), and Repeated Incremental Pruning to Produce Error Reduction (RIPPER)) with distinct computational characteristics are ensembled based on stacking and the resulting model is further enhanced using a forest penalizing attribute (FPA) model. The synthetic minority oversampling technique (SMOTE) is integrated with the proposed HMSE to balance the skewed class label present in the original experimental datasets. Extensive tests were carried out to determine the effectiveness of the proposed HMSE and S-HMSE on standard telecom CCP datasets. Observed findings from the experimental results showed that HMSE and S-HMSE can effectively predict churners even with the class imbalance (skewed datasets) problem. In addition, comparison studies demonstrated that the suggested S-HMSE offered improved prediction performance and optimum solutions for CCP in the telecom sector in comparison with baseline classifiers, homogeneous ensemble methods, and current CCP approaches.

Early detection of chronic kidney disease using recursive feature elimination and cross-validated XGBoost model

Chronic kidney disease (CKD) is listed among the top 20 leading causes of death worldwide, highlighting the urgent need for effective methods of early disease prediction for chronic conditions. This paper presents a novel strategy for the early diagnosis of CKD, leveraging machine learning (ML) methods to support researchers in their pursuit of preventative approaches. The study incorporates data from 400 patients, analyzing a comprehensive set of 25 attributes. The missing values in the dataset handled by mean and mode statistics were employed for numerical and nominal dimensions respectively. The recursive feature elimination (RFE) with cross-validation (CV) technique was utilized to identify the most crucial attributes. The resulting predictive model achieved positive outcomes in terms of accuracy, precision, recall, and [Formula: see text]-measure across all implemented classification algorithms (random forest (RF), Naive Bayes (NB), decision tree (DT), multi-layer neural network (MLNN), partial decision tree algorithm (PART), repeated incremental pruning to produce error reduction (RIPPER), [Formula: see text] Star and extreme gradient boosting (XGBoost)). Notably, the XGBoost classification algorithm demonstrated an impressive accuracy of up to 99.5%. The utilization of predictive models is indispensable for healthcare professionals, as they play a pivotal role in enabling early detection of CKD and mitigating the risk of kidney failure occurrence. This research offers valuable insights into the development of reliable methods for the early identification of CKD, providing a promising avenue for enhancing patient outcomes and reducing the burden of this prevalent chronic condition.

Distinguishing Lewy Body Dementia from Alzheimer's Disease using Machine Learning on Heterogeneous Data: A Feasibility Study.

Dementia with Lewy Bodies (DLB) is the second most common form of dementia, but diagnostic markers for DLB can be expensive and inaccessible, and many cases of DLB are undiagnosed. This work applies machine learning techniques to determine the feasibility of distinguishing DLB from Alzheimer's Disease (AD) using heterogeneous data features. The Repeated Incremental Pruning to Produce Error Reduction (RIPPER) algorithm was first applied using a Leave-One-Out Cross-Validation protocol to a dataset comprising DLB and AD cases. Then, interpretable association rule-based diagnostic classifiers were obtained for distinguishing DLB from AD. The various diagnostic classifiers generated by this process had high accuracy over the whole dataset (mean accuracy of 94%). The mean accuracy in classifying their out-of-sample case was 80.5%. Every classifier generated consisted of very simple structure, each using 1-2 classification rules and 1-3 data features. As a group, the classifiers were heterogeneous and used several different data features. In particular, some of the classifiers used very simple and inexpensive diagnostic features, yet with high diagnostic accuracy. This work suggests that opportunities may exist for incorporating accessible diagnostic assessments while improving diagnostic rate for DLB. Clinical Relevance- Simple and interpretable high-performing machine learning algorithms identified a variety of readily available clinical assessments for differential diagnosis of dementia offering the opportunities to incorporate various simple and inexpensive screening tests for DLB and addressing the problem of DLB underdiagnosis.

Identification of Type 2 Diabetes Biomarkers From Mixed Single-Cell Sequencing Data With Feature Selection Methods.

Diabetes is the most common disease and a major threat to human health. Type 2 diabetes (T2D) makes up about 90% of all cases. With the development of high-throughput sequencing technologies, more and more fundamental pathogenesis of T2D at genetic and transcriptomic levels has been revealed. The recent single-cell sequencing can further reveal the cellular heterogenicity of complex diseases in an unprecedented way. With the expectation on the molecular essence of T2D across multiple cell types, we investigated the expression profiling of more than 1,600 single cells (949 cells from T2D patients and 651 cells from normal controls) and identified the differential expression profiling and characteristics at the transcriptomics level that can distinguish such two groups of cells at the single-cell level. The expression profile was analyzed by several machine learning algorithms, including Monte Carlo feature selection, support vector machine, and repeated incremental pruning to produce error reduction (RIPPER). On one hand, some T2D-associated genes (MTND4P24, MTND2P28, and LOC100128906) were discovered. On the other hand, we revealed novel potential pathogenic mechanisms in a rule manner. They are induced by newly recognized genes and neglected by traditional bulk sequencing techniques. Particularly, the newly identified T2D genes were shown to follow specific quantitative rules with diabetes prediction potentials, and such rules further indicated several potential functional crosstalks involved in T2D.

Predicting Mortality in Patients with Stroke Using Data Mining Techniques

The mortality due to stroke is increasing. Accurate prediction of stroke-caused death is very important for healthcare. Data mining methods are novel ways to predict these mortality risks. The aim of this study is to employ popular data mining algorithms to predict the survival of stroke patients and extract decision rules. The data on stroke patients (n=4149) were collected from paper medical records. Missing data were managed using the multiple imputation method. Also, the target variable was balanced using methods such as over-sampling, under-sampling and Synthetic Minority Oversampling (SMOTE). The support vector machine (SVM), decision tree, and logistic regression (LR) algorithms were employed to predict the survival of stroke patients. Also, the Repeated Incremental Pruning to Produce Error Reduction (RIPPER) algorithm was used to extract the decision rules from the main dataset. LR outperformed other algorithms in terms of accuracy (76.96%), sensitivity (79.06%) and kappa (33.34). However, specificity (65.35%) and AUC (0.77) scores were lower than those of other algorithms. An independent dataset with 234 records was selected to challenge the LR algorithm with the best performance from the main dataset. After employing this algorithm on the external validation dataset, its performance was improved in accuracy (79.91%), sensitivity (83.94%), kappa (39.26) and AUC (0.8), but not in specificity (60.98%). The constructed model predicted the survival of stroke patients with high scores and useful rules were extracted for clinical usage.

Heterogeneous Ensemble with Combined Dimensionality Reduction for Social Spam Detection

This study presents a novel framework based on a heterogeneous ensemble method and a hybrid dimensionality reduction technique for spam detection in micro-blogging social networks. A hybrid of Information Gain (IG) and Principal Component Analysis (PCA) (dimensionality reduction) was implemented for the selection of important features and a heterogeneous ensemble consisting of Naïve Bayes (NB), K Nearest Neighbor (KNN), Logistic Regression (LR) and Repeated Incremental Pruning to Produce Error Reduction (RIPPER) classifiers based on Average of Probabilities (AOP) was used for spam detection. The proposed framework was applied on MPI_SWS and SAC’13 Tip spam datasets and the developed models were evaluated based on accuracy, precision, recall, f-measure, and area under the curve (AUC). From the experimental results, the proposed framework (that is, Ensemble + IG + PCA) outperformed other experimented methods on studied spam datasets. Specifically, the proposed method had an average accuracy value of 87.5%, an average precision score of 0.877, an average recall value of 0.845, an average F-measure value of 0.872 and an average AUC value of 0.943. Also, the proposed method had better performance than some existing methods. Consequently, this study has shown that addressing high dimensionality in spam datasets, in this case, a hybrid of IG and PCA with a heterogeneous ensemble method can produce a more effective method for detecting spam contents.

Human Papillomavirus Targeted Immunotherapy Outcome Prediction Using Machine Learning

Warts caused by the Human Papillomavirus (HPV) is a highly contagious disease, and affects several million people across the globe every year, in the form of small lesions on the skin, commonly known as warts. Warts can be treated effectively with several methods, the most effective being Immunotherapy and Cryotherapy. Our research is focused on the performance comparison of modern Machine Learning classification techniques to predict the outcome (positive or negative) of Immunotherapy treatment given to a patient, by using patient data as input features to our classifiers. The precision, recall, f-measure and accuracy were used to compare the performance of the various classifiers considered in this study. We considered Logistic Regression, ZeroR, AdaBoost, K-Nearest Neighbours (KNN), Support Vector Machines (SVM), Gradient Boosting, Repeated Incremental Pruning to Produce Error Reduction (RIPPER), Decision Trees and Random Forests. The ZeroR classifier was used as a baseline to provide us with insights into the skewed nature of the data, so as to enable us to better understand the comparison in performance of the various classifiers.

Multi-Omics Analysis of Acute Lymphoblastic Leukemia Identified the Methylation and Expression Differences Between BCP-ALL and T-ALL.

Acute lymphoblastic leukemia (ALL) as a common cancer is a heterogeneous disease which is mainly divided into BCP-ALL and T-ALL, accounting for 80–85% and 15–20%, respectively. There are many differences between BCP-ALL and T-ALL, including prognosis, treatment, drug screening, gene research and so on. In this study, starting with methylation and gene expression data, we analyzed the molecular differences between BCP-ALL and T-ALL and identified the multi-omics signatures using Boruta and Monte Carlo feature selection methods. There were 7 expression signature genes (CD3D, VPREB3, HLA-DRA, PAX5, BLNK, GALNT6, SLC4A8) and 168 methylation sites corresponding to 175 methylation signature genes. The overall accuracy, accuracy of BCP-ALL, accuracy of T-ALL of the RIPPER (Repeated Incremental Pruning to Produce Error Reduction) classifier using these signatures evaluated with 10-fold cross validation repeated 3 times were 0.973, 0.990, and 0.933, respectively. Two overlapped genes between 175 methylation signature genes and 7 expression signature genes were CD3D and VPREB3. The network analysis of the methylation and expression signature genes suggested that their common gene, CD3D, was not only different on both methylation and expression levels, but also played a key regulatory role as hub on the network. Our results provided insights of understanding the underlying molecular mechanisms of ALL and facilitated more precision diagnosis and treatment of ALL.

A New Two-step Ensemble Learning Model for Improving Stress Prediction of Automobile Drivers

Commuting when there is a significant volume of traffic congestion has been acknowledged as one of the key factors causing stress. Significant levels of stress whilst driving are seen to have a profoundly negative effect on the actions and ability of a driver; this has the capacity to result in risks, hazards and accidents. As such, there is a recognized need to determine drivers’ levels of stress and accordingly predict the key causes responsible for high levels of stress. In this work, the objective is centred on providing an ensemble machine learning framework in order to determine the stress levels of drivers. Moreover, the study also provides a fresh set of data, as gathered from 14 different drivers, with data collection having taken place during driving in Amman, Jordan. Data was gathered via the implementation of a wearable biomedical instrument that was attached to the driver on a continuous basis in order to gather physiological data. The data gathered was accordingly categorised into two different groups: ‘Yes’, which represents the presence of stress, whilst ‘No’ represents the absence of stress. Importantly, in an effort to circumvent the negative impact of driver instances with a minority class on stress predictions, oversampling technique was applied. A two-step ensemble classifier was developed through bringing together the findings from random forest, decision tree, and Repeated Incremental Pruning to Produce Error Reduction (RIPPER) classifiers, which was then inputted into a Multi-Layer Perceptron neural network. The experimental findings highlight that the suggested framework is far more precise and has a more scalable capacity when compared with all classifiers in relation to accuracy, g-mean measures and sensitivity.

Early Prediction of a Team Performance in the Initial Assessment Phases of a Software Project for Sustainable Software Engineering Education

Software engineering is a competitive field in education and practice. Software projects are key elements of software engineering courses. Software projects feature a fusion of process and product. The process reflects the methodology of performing the overall software engineering practice. The software product is the final product produced by applying the process. Like any other academic domain, an early evaluation of the software product being developed is vital to identify the at-risk teams for sustainable education in software engineering. Guidance and instructor attention can help overcome the confusion and difficulties of low performing teams. This study proposed a hybrid approach of information gain feature selection with a J48 decision tree to predict the earliest possible phase for final performance prediction. The proposed technique was compared with the state-of-the-art machine learning (ML) classifiers, naïve Bayes (NB), artificial neural network (ANN), logistic regression (LR), simple logistic regression (SLR), repeated incremental pruning to produce error reduction (RIPPER), and sequential minimal optimization (SMO). The goal of this process is to predict the teams expected to obtain a below-average grade in software product development. The proposed technique outperforms others in the prediction of low performing teams at an early assessment stage. The proposed J48-based technique outperforms others by making 89% correct predictions.

Investigating the gene expression profiles of cells in seven embryonic stages with machine learning algorithms

The development of embryonic cells involves several continuous stages, and some genes are related to embryogenesis. To date, few studies have systematically investigated changes in gene expression profiles during mammalian embryogenesis. In this study, a computational analysis using machine learning algorithms was performed on the gene expression profiles of mouse embryonic cells at seven stages. First, the profiles were analyzed through a powerful Monte Carlo feature selection method for the generation of a feature list. Second, increment feature selection was applied on the list by incorporating two classification algorithms: support vector machine (SVM) and repeated incremental pruning to produce error reduction (RIPPER). Through SVM, we extracted several latent gene biomarkers, indicating the stages of embryonic cells, and constructed an optimal SVM classifier that produced a nearly perfect classification of embryonic cells. Furthermore, some interesting rules were accessed by the RIPPER algorithm, suggesting different expression patterns for different stages.

Machine-tool condition monitoring with Gaussian mixture models-based dynamic probabilistic clustering

The combination of artificial intelligence with data, computing power, and new algorithms can provide important tools for solving engineering problems, such as machine-tool condition monitoring. However, many of these problems require algorithms that can perform in highly dynamic scenarios where the data streams have extremely high sampling rates from different types of variables. The unsupervised learning algorithm based on Gaussian mixture models called Gaussian-based dynamic probabilistic clustering (GDPC) is one of these tools. However, this algorithm may have major limitations if a large amount of concept drifts associated with transients occurs within the data stream. GDPC becomes unstable under these conditions, so we propose a new algorithm called GDPC+ to increase its robustness. GDPC+ represents an important improvement because we introduce: (a) automatic selection of the number of mixture components based on the Bayesian information criterion (BIC), and (b) concept drift transition stabilization based on Cauchy–Schwarz divergence integrated with the Dickey–Fuller test. Thus, GDPC+ can perform better in highly dynamic scenarios than GDPC in terms of the number of false positives. The behavior of GDPC+ was investigated using random synthetic data streams and in a real data stream-based condition monitoring obtained from a machine-tool that produces engine crankshafts at high speed. We found that the initial temporal window size can be used to adapt the algorithm to different analytical requirements. The clustering results were also investigated by induction of the rules generated by the repeated incremental pruning to produce error reduction (RIPPER) algorithm in order to provide insights from the underlying monitored process and its associated concept drifts.

Longitudinal Analysis of Tactical Strategy in the Men's Division of the Ultimate Fighting Championship.

This study explored longitudinal changes in contemporary mixed martial arts (MMA) combat within the Ultimate Fighting Championship (UFC). A secondary aim was to investigate how bout duration influences the contribution of performance indicators on outcome. Data were acquired via the official analytics provider to the UFC (FightMetric). Male fights with a winner from between 2000 and 2015 (n = 2,831) were examined, with 13 common performance indicators attained during each round for each participant along with duration (min) and year of fight. Non-metric dimensional scaling (nMDS) was used to examine bout characteristics by year. The Repeated Incremental Pruning to Produce Error Reduction (RIPPER) algorithm was run to determine a set of rules to explain bout outcome. The nMDS displayed that winning bout performance indicator attributes were dissimilar across the years. Eight rules were generated from the RIPPER, with fight duration featuring in three of eight rules. Distinct shifts occurred (albeit without linear trend) in performance indicator characteristics during the observed period. This was characterized by a more diverse combat style in the years following 2008. However, offensive grappling has remained a key factor regardless of year, and is influenced by bout duration.

Multi-Objective Optimization Energy Management Strategy for Fuel Cell Hybrid Vehicles Based on Rule Learning

In this paper, a rule learning-based energy management strategy is proposed to achieve optimal energy consumption economy and prolong the batteries lifetime for a fuel cell hybrid electric vehicle (FCHEV). First, the Pontryagin’s minimum principle (PMP) is used to obtain the optimal global solution. Then, K-means algorithm is adopted to simplify the optimal database composed of the optimized data based on PMP and the corresponding driving cycle features. According to the simplification data set, the improved repeated incremental pruning to produce error reduction (RIPPER) algorithm based on rule learning theory is used to learn the rules. Finally, the multiple linear regression algorithms are utilized to fit the data in the rule set. Simulation results validate that the proposed strategy can achieve more than 94% of the PMP strategy in the fuel consumption and also can prolong the batteries lifetime.

Identification of leukemia stem cell expression signatures through Monte Carlo feature selection strategy and support vector machine.

Acute myeloid leukemia (AML) is a type of blood cancer characterized by the rapid growth of immature white blood cells from the bone marrow. Therapy resistance resulting from the persistence of leukemia stem cells (LSCs) are found in numerous patients. Comparative transcriptome studies have been previously conducted to analyze differentially expressed genes between LSC+ and LSC- cells. However, these studies mainly focused on a limited number of genes with the most obvious expression differences between the two cell types. We developed a computational approach incorporating several machine learning algorithms, including Monte Carlo feature selection (MCFS), incremental feature selection (IFS), support vector machine (SVM), Repeated Incremental Pruning to Produce Error Reduction (RIPPER), to identify gene expression features specific to LSCs. One thousand 0ne hudred fifty-nine features (genes) were first identified, which can be used to build the optimal SVM classifier for distinguishing LSC+ and LSC- cells. Among these 1159 genes, the top 17 genes were identified as LSC-specific biomarkers. In addition, six classification rules were produced by RIPPER algorithm. The subsequent literature review on these features/genes and the classification rules and functional enrichment analyses of the 1159 features/genes confirmed the relevance of extracted genes and rules to the characteristics of LSCs.

Analysis of Expression Pattern of snoRNAs in Different Cancer Types with Machine Learning Algorithms.

Small nucleolar RNAs (snoRNAs) are a new type of functional small RNAs involved in the chemical modifications of rRNAs, tRNAs, and small nuclear RNAs. It is reported that they play important roles in tumorigenesis via various regulatory modes. snoRNAs can both participate in the regulation of methylation and pseudouridylation and regulate the expression pattern of their host genes. This research investigated the expression pattern of snoRNAs in eight major cancer types in TCGA via several machine learning algorithms. The expression levels of snoRNAs were first analyzed by a powerful feature selection method, Monte Carlo feature selection (MCFS). A feature list and some informative features were accessed. Then, the incremental feature selection (IFS) was applied to the feature list to extract optimal features/snoRNAs, which can make the support vector machine (SVM) yield best performance. The discriminative snoRNAs included HBII-52-14, HBII-336, SNORD123, HBII-85-29, HBII-420, U3, HBI-43, SNORD116, SNORA73B, SCARNA4, HBII-85-20, etc., on which the SVM can provide a Matthew’s correlation coefficient (MCC) of 0.881 for predicting these eight cancer types. On the other hand, the informative features were fed into the Johnson reducer and repeated incremental pruning to produce error reduction (RIPPER) algorithms to generate classification rules, which can clearly show different snoRNAs expression patterns in different cancer types. The analysis results indicated that extracted discriminative snoRNAs can be important for identifying cancer samples in different types and the expression pattern of snoRNAs in different cancer types can be partly uncovered by quantitative recognition rules.

Evaluation of Plasma Extracellular Vesicle MicroRNA Signatures for Lung Adenocarcinoma and Granuloma With Monte-Carlo Feature Selection Method.

Extracellular Vesicle (EV) is a compilation of secreted vesicles, including micro vesicles, large oncosomes, and exosomes. It can be used in non-invasive diagnosis. MicroRNAs (miRNAs) processed by exosomes can be detected by liquid biopsy. To objectively evaluate the discriminative ability of miRNAs from whole plasma, EV and EV-free plasma, we analyzed the miRNA expression profiles in whole plasma, EV and EV-free plasma of 10 lung adenocarcinoma and 9 granuloma patients. With Monte-Carlo feature selection method, the top discriminative miRNAs in whole plasma, EV and EV-free plasma were identified, and they were quite different. Using the Repeated Incremental Pruning to Produce Error Reduction (RIPPER) method, we learned the classification rules: in whole plasma, granuloma patients did not express hsa-miR-223-3p while the lung adenocarcinoma patients expressed hsa-miR-223-3p; in EV, the hsa-miR-23b-3p was highly expressed in granuloma patients but not lung adenocarcinoma patients; in EV-free plasma, hsa-miR-376a-3p was expressed in granuloma patients but barely expressed in lung adenocarcinoma patients. For prediction performance, whole plasma had the highest weighted accuracy and EV outperformed EV-free plasma. Our results suggested that EV can be used as lung cancer biomarker. However, since it is less stable and not easy to detect, there are still technological difficulties to overcome.

A Two-Step Knowledge Extraction Framework for Improving Disease Diagnosis

AbstractIn the last decades, various methodologies have been proposed by the researchers for developing effective disease diagnosis support systems (DDSSs). The present research proposes a two-step framework in which an entropy-based feature-selection approach is introduced in the first step and a rule-base hybrid model using Perfect Rule Induction by Sequential Method (PRISM) is explored in the subsequent step for effective diagnosis of diseases. The suggested feature-selection technique is validated using five state-of-the-art classifiers namely C4.5 (a decision tree-based classifier), naïve Bayes (NB), Repeated Incremental Pruning to Produce Error Reduction (RIPPER), neural network (NN) and support vector machine (SVM) over fourteen benchmark diseases that are very common and the leading causes of deaths. Next, on the basis of top three performance metrics, viz., prediction accuracy, sensitivity and false positive rate, the performance of the hybrid model over the datasets is compared with its base learner: PRISM, two other competent learners namely C4.5 and NN, and some specialized models. The empirical outcomes positively demonstrate that the hybrid model with application of feature-selection method is a generic model and effective in diagnosing diseases. More importantly, the model not only is able to produce good results but also to elucidate its knowledge in understandable: IF-THEN form (convenient for medical practitioners).

Application of FURIA for Finding the Faults in a Hydraulic Brake System Using a Vibration Analysis through a Machine Learning Approach

Vibration-based continuous monitoring system for fault diagnosis of automobile hydraulic brake system is presented in this study. This study uses a machine learning approach for the fault diagnosis study. A hydraulic brake system test rig was fabricated. The vibration signals were acquired from the brake system under different simulated fault conditions using a piezoelectric transducer. The histogram features were extracted from the acquired vibration signals. The feature selection process was carried out using a decision tree. The selected features were classified using fuzzy unordered rule induction algorithm ( FURIA ) and Repeated Incremental Pruning to Produce Error Reduction ( RIPPER ) algorithm. The classification results of both algorithms for fault diagnosis of a hydraulic brake system were presented. Compared to RIPPER and J48 decision tree, the FURIA performs better and produced 98.73 % as the classification accuracy.

Identifying Essential Signature Genes and Expression Rules Associated With Distinctive Development Stages of Early Embryonic Cells

An embryo develops from a single-celled zygote, which produces a multi-cellular organism by mitosis. Due to the complication of processes and mechanisms, research on embryo cell clusters in different early embryo developmental stages with significant phenotypic differences is still lacking. In this work, we identified some gene characters and expression rules to classify these individual cells using several advanced computational methods. The single cell expression profiles of embryo cells were analyzed by the Monte Carlo feature selection (MCFS) method, resulting in a feature list. Then, the incremental feature selection (IFS) method, incorporating support vector machine (SVM), applied on such list to extract key gene characters. These gene characters include KHDC1, HMGN1, DCP, GDF9, RNF11, DNMT3L, and CDX1. Furthermore, a rule learning algorithm, Repeated Incremental Pruning to Produce Error Reduction (RIPPER), was applied to the informative features yielded by MCFS method, producing a group of classification rules. These rules can clearly uncover different expression patterns on cells in different stages. This study provided a group of effective gene signatures and rules for embryo cell subtyping and presented an applicable computational tool to further dig into the regulatory mechanisms of embryo development.