A systematic review of machine learning techniques for software fault prediction

BackgroundSoftware fault prediction is the process of developing models that can be used by the software practitioners in the early phases of software development life cycle for detecting faulty constructs such as modules or classes. There are various machine learning techniques used in the past for predicting faults.MethodIn this study we perform a systematic review of studies from January 1991 to October 2013 in the literature that use the machine learning techniques for software fault prediction. We assess the performance capability of the machine learning techniques in existing research for software fault prediction. We also compare the performance of the machine learning techniques with the statistical techniques and other machine learning techniques. Further the strengths and weaknesses of machine learning techniques are summarized.ResultsIn this paper we have identified 64 primary studies and seven categories of the machine learning techniques. The results prove the prediction capability of the machine learning techniques for classifying module/class as fault prone or not fault prone. The models using the machine learning techniques for estimating software fault proneness outperform the traditional statistical models.ConclusionBased on the results obtained from the systematic review, we conclude that the machine learning techniques have the ability for predicting software fault proneness and can be used by software practitioners and researchers. However, the application of the machine learning techniques in software fault prediction is still limited and more number of studies should be carried out in order to obtain well formed and generalizable results. We provide future guidelines to practitioners and researchers based on the results obtained in this work.     

Examining Factors That Influence the Effectiveness of Learning Objects in Mathematics Classrooms

Abstract

Learning objects are interactive online tools that support the acquisition of specific concepts. Limited research has been conducted on factors that affect the use of learning objects in K–12 mathematics classrooms. The current study examines the influence of student characteristics (gender, age, computer comfort level, subject comfort level, and mathematics grade), instructional design (structured vs. open ended), and teaching strategy (teacher led vs. student based) on student attitudes toward the use of learning objects and learning performance. Data in the form of surveys and pre- and posttests were collected from 286 middle and secondary school students. Higher computer and subject area comfort ratings were significantly correlated with more positive student attitudes about learning objects. Older students in higher grades learned more than younger students in lower grades after using learning objects. Learning performance was significantly higher for students who used structured (vs. open-ended) learning objects and participated in teacher-led (vs. student-based) lessons. It is speculated that younger students might need more scaffolding when using mathematics-based learning objects.     

Prediction of high-strength concrete: high-order response surface methodology modeling approach

In the concrete industry, compressive strength is the most essential mechanical property. Therefore, insufficient compressive strength may lead to dangerous failure and, thus, becomes very difficult to repair. Consequently, early, and precise prediction of concrete strength is a major issue facing researchers and concrete designers. In this study, high-order response surface methodology (HORSM) is used to develop a prediction model to accurately predict the compressive strength of high-strength concrete (HSC). Different polynomial degrees order ranging from 2 to 5 is used in this model. The HORSM, with five-order polynomial degree, model outperforms several artificial intelligence (AI) modeling approaches which are carried out widely in the prediction of HSC compression strength. Besides, support vector machine (SVM) model was developed in this study and compared with the HORSM. The HORSM models outperformed the SVM models according to different statistical measures. Additionally, HORSM models managed to perfectly predict the HSC compressive strength in less than one second to accomplish the learning processes. While, other AI models including SVM much longer time. Lastly, the use of HORSM for the first time in the concrete technology field provided much accurate prediction results and it has great potential in the field of concrete technology.

     

A statistical model including age to predict passenger postures in the rear seats of automobiles

Few statistical models of rear seat passenger posture have been published, and none has taken into account the effects of occupant age. This study developed new statistical models for predicting passenger postures in the rear seats of automobiles. Postures of 89 adults with a wide range of age and body size were measured in a laboratory mock-up in seven seat configurations. Posture-prediction models for female and male passengers were separately developed by stepwise regression using age, body dimensions, seat configurations and two-way interactions as potential predictors. Passenger posture was significantly associated with age and the effects of other two-way interaction variables depended on age. A set of posture-prediction models are presented for women and men, and the prediction results are compared with previously published models. This study is the first study of passenger posture to include a large cohort of older passengers and the first to report a significant effect of age for adults. The presented models can be used to position computational and physical human models for vehicle design and assessment.

Practitioner Summary:

The significant effects of age, body dimensions and seat configuration on rear seat passenger posture were identified. The models can be used to accurately position computational human models or crash test dummies for older passengers in known rear seat configurations.     

Examining endplate fatigue failure during cyclic compression loading with variable and consistent peak magnitudes using a force weighting adjustment approach: an in vitro study

Abstract

Repetitive movement is common in many occupational contexts. Therefore, cumulative load is a widely recognised risk factor for lowback injury. This study quantified the effect of force weighting factors on cumulative load estimates and injury prediction during cyclic loading. Forty-eight porcine cervical spine motion segments were assigned to experimental groups that differed by average peak compression magnitude (30%, 50% and 70% of predicted tolerance) and amplitude variation (consistent, variable). Cyclic loading was performed at a frequency of 0.5?Hz until fatigue failure occurred. Weighting factors were determined and applied instantaneously. Inclusion of weighting factors resulted in statistically similar cumulative load estimates at injury between variable and consistent loading (p?>?.071). Further, survivorship was generally greater when the peak compression magnitude was consistent compared to variable. These results emphasise the importance of weighting factors as an equalisation tool for the evaluation of cumulative low back loading exposures in occupational contexts.

Practitioner summary: Weighting factors can equalise the risk of injury based on compression magnitude. When weighted, the cumulative compression was similar between consistent and variable cyclic loading protocols, despite being significantly different when unweighted and having similar injury rates. Therefore, assessing representative occupational exposures without evaluating task performance variability may underestimate injury risk.

Abbreviations: FSU: functional spinal unit; UCT: ultimate compression tolerance     

自纠正噪声标签的人脸美丽预测

目的 人脸美丽预测是研究如何使计算机具有与人类相似的人脸美丽判断或预测能力,然而利用深度神经网络进行人脸美丽预测存在过度拟合噪声标签样本问题,从而影响深度神经网络的泛化性。因此,本文提出一种自纠正噪声标签方法用于人脸美丽预测。方法 该方法包括自训练教师模型机制和重标签再训练机制。自训练教师模型机制以自训练的方式获得教师模型,帮助学生模型进行干净样本选择和训练,直至学生模型泛化能力超过教师模型并成为新的教师模型,并不断重复该过程;重标签再训练机制通过比较最大预测概率和标签对应预测概率,从而纠正噪声标签。同时,利用纠正后的数据反复执行自训练教师模型机制。结果 在大规模人脸美丽数据库LSFBD (large scale facial beauty database)和SCUT-FBP5500数据库上进行实验。结果表明,本文方法在人工合成噪声标签的条件下可降低噪声标签的负面影响,同时在原始LSFBD数据库和SCUT-FBP5500数据库上分别取得60.8%和75.5%的准确率,高于常规方法。结论 在人工合成噪声标签条件下的LSFBD和SCUT-FBP5500数据库以及原始LSFBD和SCUT-FBP5500数据库上的实验表明,所提自纠正噪声标签方法具有选择干净样本学习、充分利用全部数据的特点,可降低噪声标签的负面影响,能在一定程度上降低人脸美丽预测中噪声标签的负面影响,提高预测准确率。     

“Extra-Ordinary” Ergonomics: How to Accommodate Small and Big Persons,the Disabled and Elderly,Expectant Mothers,and Children

Abstract

Data from a previous study of soldier driving postures and seating positions were analysed to develop statistical models for defining accommodation of driver seating positions in military vehicles. Regression models were created for seating accommodation applicable to driver positions with a fixed heel point and a range of steering wheel locations in typical tactical vehicles. The models predict the driver-selected seat position as a function of population anthropometry and vehicle layout. These models are the first driver accommodation models considering the effects of body armor and body-borne gear. The obtained results can benefit the design of military vehicles, and the methods can also be extended to be utilised in the development of seating accommodation models for other driving environments where protective equipment affects driver seating posture, such as vehicles used by law-enforcement officers and firefighters.

Practitioner Summary: A large-scale laboratory study of soldier driving posture and seating position was designed to focus on tactical vehicle (truck) designs. Regression techniques are utilised to develop accommodation models suitable for tactical vehicles. These are the first seating accommodation models based on soldier data to consider the effects of personal protective equipment and body-borne gear.     

Simulation and prediction of hydrological processes based on firefly algorithm with deep learning and support vector for regression

ABSTRACT

Hydrological processes are hard to accurately simulate and predict because of various natural and human influences. In order to improve the simulation and prediction accuracy of the hydrological process, the firefly algorithm with deep learning (DLFA) was used in this study to optimise the parameters of support vector for regression (SVR) automatically, and a prediction model was established based on DLFA and SVR. The hydrological process of Huangfuchuan in Fugu County, Shanxi Province was taken as the research object to verify the performance of the prediction model, and the results were compared with those by the other six prediction models. The experimental results showed that the proposed prediction model achieved improved prediction performance compared with the other six models.     

Faculty Begin to Replace Textbooks with “Freely” Accessible Online Resources

ABSTRACT

More and more students are financially unable to acquire, or deliberately choose to go without course textbooks. A variety of commercial and noncommercial initiatives have materialized to address the student success challenge of learning material access inequality in the classroom. There is a gap between how higher education faculty plan to teach a course and the actual learning environment that exists in practice. Faculty are beginning to experiment with freely available and licensed library materials as a substitute for commercial textbooks and course packages to address the failure of textbook publishers to reach a price point that entices students to buy textbooks. The results thus far are promising. Some courses can be delivered today using only “freely available” learning resources, some using a mix of fee based and free, while others cannot be delivered using any freely available resources at all due to a lack of availability.     

Transition Between University Students to Teachers: Practice in the Middle

Abstract

This article presents a study done in an elementary mathematics methods course that focused on the transition of novice teachers’ epistemological stances: former elementary student, university student, and teacher stances. In order to help them develop the teacher stance, we designed a three-phase activity, where two phases took place inside class and the last one occurred outside of class. Novice teachers were given an assignment where they had to rehearse a count in class and enact it in front of a small group of students. They had to write reflections on their rehearsal and enactment. Interviews were done 4 months after the end of the course. The results show that the reflections about mathematics in relation to the use of new teaching practices on eliciting students’ thinking allowed the novice teachers to develop the teacher stance.     

Modeling of arsenic, chromium and cadmium removal by nanofiltration process using genetic programming

In this paper, genetic programming (GP) as a novel approach for the explicit formulation of nanofiltration (NF) process performance is presented. The objective of this study is to develop robust models based on experimental data for prediction the membrane rejection of arsenic, chromium and cadmium ions in a NF pilot-scale system using GP. Feed concentration and transmembrane pressure were considered as input parameters of the models. The ions rejection is considered as output parameter of the models. Some statistical parameters were considered and calculated in order to investigate the reliability of each model. The results showed quite satisfactory accuracies of the proposed models based on GP. The results also nominated GP as a potential tool for identifying the behavior of a NF system.     

Strength evaluation of granite block samples with different predictive models

Over the last decade, application of soft computing techniques has rapidly grown up in different scientific fields, especially in rock mechanics. One of these cases relates to indirect assessment of uniaxial compressive strength (UCS) of rock samples with different artificial intelligent-based methods. In fact, the main advantage of such systems is to readily remove some difficulties arising in direct assessment of UCS, such as time-consuming and costly UCS test procedure. This study puts an effort to propose four accurate and practical predictive models of UCS using artificial neural network (ANN), hybrid ANN with imperialism competitive algorithm (ICA–ANN), hybrid ANN with artificial bee colony (ABC–ANN) and genetic programming (GP) approaches. To reach the aim of the current study, an experimental database containing a total of 71 data sets was set up by performing a number of laboratory tests on the rock samples collected from a tunnel site in Malaysia. To construct the desired predictive models of UCS based on training and test patterns, a combination of several rock characteristics with the most influence on UCS has been used as input parameters, i.e. porosity (n), Schmidt hammer rebound number (R), p-wave velocity (V_p) and point load strength index (Is(50)). To evaluate and compare the prediction precision of the developed models, a series of statistical indices, such as root mean squared error (RMSE), determination coefficient (R²) and variance account for (VAF) are utilized. Based on the simulation results and the measured indices, it was observed that the proposed GP model with the training and test RMSE values 0.0726 and 0.0691, respectively, gives better performance as compared to the other proposed models with values of (0.0740 and 0.0885), (0.0785 and 0.0742), and (0.0746 and 0.0771) for ANN, ICA–ANN and ABC–ANN, respectively. Moreover, a parametric analysis is accomplished on the proposed GP model to further verify its generalization capability. Hence, this GP-based model can be considered as a new applicable equation to accurately estimate the uniaxial compressive strength of granite block samples.

     

Modelling the hare and the tortoise: predicting the range of in-vehicle task times using critical path analysis

Analytic models can enable predictions about important aspects of the usability of in-vehicle information systems (IVIS) to be made at an early stage of the product development process. Task times provide a quantitative measure of user performance and are therefore important in the evaluation of IVIS usability. In this study, critical path analysis (CPA) was used to model IVIS task times in a stationary vehicle, and the technique was extended to produce predictions for slowperson and fastperson performance, as well as average user (middleperson) performance. The CPA-predicted task times were compared to task times recorded in an empirical simulator study of IVIS interaction, and the predicted times were, on average, within acceptable precision limits. This work forms the foundation for extension of the CPA model to predict IVIS task times in a moving vehicle, to reflect the demands of the dual-task driving scenario.

Practitioner Summary: The CPA method was extended for the prediction of slowperson and fastperson IVIS task times. Comparison of the model predictions with empirical data demonstrated acceptable precision. The CPA model can be used in early IVIS evaluation; however, there is a need to extend it to represent the dual-task driving scenario.     

Information Literacy and the Distant Student

SUMMARY

This article discusses the development and implementation of a required, credit bearing online information literacy course at the University of Maryland University College. Key factors in its success include administrative support, student and faculty interaction in the online classroom, and outcomes assessment. Student persistence in the course is high, and grade distributions indicate that students are being challenged.     

Prediction of tropospheric ozone concentrations: Application of a methodology based on the Darwin’s Theory of Evolution

This study aims to predict the next day hourly average tropospheric ozone (O₃) concentrations using genetic programming (GP). Due to the complexity of this problem, GP is an adequate methodology as it can optimize, simultaneously, the structure of the model and its parameters. It is an artificial intelligence methodology that uses the same principles of the Darwinian Theory of Evolution. GP enables the automatic generation of mathematical expressions that are modified following an iterative process applying genetic operations.The inputs of the models were the hourly average concentrations of carbon monoxide (CO), nitrogen oxide (NO), nitrogen dioxide (NO₂) and O₃, and some meteorological variables (temperature – T; solar radiation – SR; relative humidity – RH; and wind speed – WS) measured 24 h before. GP was also applied to the principal components (PC) obtained from these variables. The analysed period was from May to July 2004 divided in training and test periods.GP was able to select the most relevant variables for prediction of O₃ concentrations. The original variables, T, RH and O₃ measured 24 h before were considered significant inputs for prediction. The selected PC had also important contributions of the same variables and of NO₂. GP models using the original variables presented better performance in training period and worse performance in test period when compared with the models obtained using PC. The results achieved using the GP methodology demonstrated that it can be very useful to solve several environmental complex problems.     

Posture normalisation of 3D body scans

For product developers that design near-body products, virtual mannequins that represent realistic body shapes, are valuable tools. With statistical shape modelling, the variability of such body shapes can be described. Shape variation captured by statistical shape models (SSMs) is often polluted by posture variations, leading to less compact models. In this paper, we propose a framework that has low computational complexity to build a posture invariant SSM, by capturing and correcting the posture of an instance. The posture-normalised SSM is shown to be substantially more compact than the non-posture-normalised SSM.

Practitioner summary: Statistical shape modelling is a technique to map out the variability of (body) shapes. This variability is often polluted by variations in posture. In this paper, we propose a framework to build a posture invariant statistical shape model.

Abbreviations: SSM: statistical shape model; 1D: one-dimensional; 3D: three-dimensional; DHM: digital human model; LBS: linear blend skinning; PCA: princial component analysis; PC: principal component; TTR: thumb tip reach.     

Neuro-genetic,neuro-imperialism and genetic programing models in predicting ultimate bearing capacity of pile

Prediction of pile-bearing capacity developing artificial intelligence models has been done over the last decade. Such predictive tools can assist geotechnical engineers to easily determine the ultimate pile bearing capacity instead of conducting any difficult field tests. The main aim of this study is to predict the bearing capacity of pile developing several smart models, i.e., neuro-genetic, neuro-imperialism, genetic programing (GP) and artificial neural network (ANN). For this purpose, a number of concrete pile characteristics and its dynamic load test specifications were investigated to select pile cross-sectional area, pile length, pile set, hammer weight and drop height as five input variables which have the most impacts on pile bearing capacity as the single output variable. It should be noted that all the aforementioned parameters were measured by conducting a series of pile driving analyzer tests on precast concrete piles located in Pekanbaru, Indonesia. The recorded data were used to establish a database of 50 test cases. With regard to data modelling, many smart models of neuro-genetic, neuro-imperialism, GP and ANN were developed and then evaluated based on the three most common statistical indices, i.e., root mean squared error (RMSE), coefficient determination (R²) and variance account for (VAF). Based on the simulation results and the computed indices’ values, it is observed that the proposed GP model with training and test RMSE values of 0.041 and 0.040, respectively, performs noticeably better than the proposed neuro-genetic model with RMSE values of 0.042 and 0.040, neuro-imperialism model with RMSE values of 0.045 and 0.059, and ANN model with RMSE values of 0.116 and 0.108 for training and test sets, respectively. Therefore, this GP-based model can provide a new applicable equation to effectively predict the ultimate pile bearing capacity.

     

Predicting discharge coefficient of triangular labyrinth weir using extreme learning machine,artificial neural network and genetic programming

Weirs are a type of hydraulic structure used to direct and transfer water flows in the canals and overflows in the dams. The important index in computing flow discharge over the weir is discharge coefficient (C _d). The aim of this study is accurate determination of the C _d in triangular labyrinth side weirs by applying three intelligence models [i.e., artificial neural network (ANN), genetic programming (GP) and extreme learning machine (ELM)]. The calculated discharge coefficients were then compared with some experimental results. In order to examine the accuracy of C _d predictions by ANN, GP and ELM methods, five statistical indices including coefficient of determination (R ²), root-mean-square error (RMSE), mean absolute percentage error (MAPE), SI and δ have been used. Results showed that R ² values in the ELM, ANN and GP methods were 0.993, 0.886 and 0.884, respectively, at training stage and 0.971, 0.965 and 0.963, respectively, at test stage. The ELM method, having MAPE, RMSE, SI and δ values of 0.81, 0.0059, 0.0082 and 0.81, respectively, at the training stage and 0.89, 0.0063, 0.0089 and 0.88, respectively, at the test stage, was superior to ANN and GP methods. The ANN model ranked next to the ELM model.