Intelligent churn prediction in telecom: employing mRMR feature selection and RotBoost based ensemble classification

Churn prediction in telecom has recently gained substantial interest of stakeholders because of associated revenue losses. Predicting telecom churners, is a challenging problem due to the enormous nature of the telecom datasets. In this regard, we propose an intelligent churn prediction system for telecom by employing efficient feature extraction technique and ensemble method. We have used Random Forest, Rotation Forest, RotBoost and DECORATE ensembles in combination with minimum redundancy and maximum relevance (mRMR), Fisher’s ratio and F-score methods to model the telecom churn prediction problem. We have observed that mRMR method returns most explanatory features compared to Fisher’s ratio and F-score, which significantly reduces the computations and help ensembles in attaining improved performance. In comparison to Random Forest, Rotation Forest and DECORATE, RotBoost in combination with mRMR features attains better prediction performance on the standard telecom datasets. The better performance of RotBoost ensemble is largely attributed to the rotation of feature space, which enables the base classifier to learn different aspects of the churners and non-churners. Moreover, the Adaboosting process in RotBoost also contributes in achieving higher prediction accuracy by handling hard instances. The performance evaluation is conducted on standard telecom datasets using AUC, sensitivity and specificity based measures. Simulation results reveal that the proposed approach based on RotBoost in combination with mRMR features (CP-MRB) is effective in handling high dimensionality of the telecom datasets. CP-MRB offers higher accuracy in predicting churners and thus is quite prospective in modeling the challenging problems of customer churn prediction in telecom.     

Detection of Circulating Tumor Cells Using the Attune NxT

Circulating tumor cells (CTCs) have been detected in many patients with different solid malignancies. It has been reported that presence of CTCs correlates with worse survival in patients with multiple types of cancer. Several techniques have been developed to detect CTCs in liquid biopsies. Currently, the only method for CTC detection that is approved by the Food and Drug Administration is CellSearch. Due to low abundance of CTCs in certain cancer types and in early stages of disease, its clinical application is currently limited to metastatic colorectal cancer, breast cancer and prostate cancer. Therefore, we aimed to develop a new method for the detection of CTCs using the Attune NxT—a flow cytometry-based application that was specifically developed to detect rare events in biological samples without the need for enrichment. When healthy donor blood samples were spiked with variable amounts of different EpCAM+EGFR+ tumor cell lines, recovery yield was on average 75%. The detection range was between 1000 and 10 cells per sample. Cell morphology was confirmed with the Attune CytPix. Analysis of blood samples from metastatic colorectal cancer patients, as well as lung cancer patients, demonstrated that increased EpCAM+EGFR+ events were detected in more than half of the patient samples. However, most of these cells showed no (tumor) cell-like morphology. Notably, CellSearch analysis of blood samples from a subset of colorectal cancer patients did not detect CTCs either, suggesting that these blood samples were negative for CTCs. Therefore, we anticipate that the Attune NxT is not superior to CellSearch in detection of low amounts of CTCs, although handling and analysis of samples is easier. Moreover, morphological confirmation is essential to distinguish between CTCs and false positive events.     

Influence of Bone Substitutes on Mesenchymal Stromal Cells in an Inflammatory Microenvironment

Bone regeneration is driven by mesenchymal stromal cells (MSCs) via their interactions with immune cells, such as macrophages (MPs). Bone substitutes, e.g., bi-calcium phosphates (BCPs), are commonly used to treat bone defects. However, little research has focused on MSC responses to BCPs in the context of inflammation. The objective of this study was to investigate whether BCPs influence MSC responses and MSC–MP interactions, at the gene and protein levels, in an inflammatory microenvironment. In setup A, human bone marrow MSCs combined with two different BCP granules (BCP 60/40 or BCP 20/80) were cultured with or without cytokine stimulation (IL1β + TNFα) to mimic acute inflammation. In setup B, U937 cell-line-derived MPs were introduced via transwell cocultures to setup A. Monolayer MSCs with and without cytokine stimulation served as controls. After 72 h, the expressions of genes related to osteogenesis, healing, inflammation and remodeling were assessed in the MSCs via quantitative polymerase chain reactions. Additionally, MSC-secreted cytokines related to healing, inflammation and chemotaxis were assessed via multiplex immunoassays. Overall, the results indicate that, under both inflammatory and non-inflammatory conditions, the BCP granules significantly regulated the MSC gene expressions towards a pro-healing genotype but had relatively little effect on the MSC secretory profiles. In the presence of the MPs (coculture), the BCPs positively regulated both the gene expression and cytokine secretion of the MSCs. Overall, similar trends in MSC responses were observed with BCP 60/40 and BCP 20/80. In summary, within the limits of in vitro models, these findings suggest that the presence of BCP granules at a surgical site may not necessarily have a detrimental effect on MSC-mediated wound healing, even in the event of inflammation.     

Performance Analysis of SDN/OpenFlow Controllers: POX Versus Floodlight

Software-Defined Networking (SDN) is an emerging network architecture that is adaptable, dynamic, cost-effective, and manageable. The SDN architecture is a form of network virtualization where the network controlling functions and forwarding functions are decoupled. A setup and configuration task of a control plane to work as an SDN controller is explained in this paper. This paper includes a brief survey of different SDN based OpenFlow-enabled controllers available in various programmable languages. This paper mainly focuses on two OpenFlow-enabled controllers, namely, POX—a Python-based controller and Floodlight—a Java-based controller. A performance comparison of both controllers is tested over different network topologies by analyzing network throughput and round-trip delay using an efficient network simulator called Mininet. A single, linear, tree and custom (user-defined) topologies are designed in Mininet by enabling external controllers. It is obtained that, a percentage improvement in round-trip time for Floodlight over POX is 11.5, 13.9, 19.6 and 14.4% for single, linear, tree and custom topology respectively. Similarly, a percentage improvement in throughput for Floodlight over POX is 5.4, 8.9, 3.8 and 4.9% for single, linear, tree and custom topology respectively.     

Expression of keratin 13, 14 and 19 in oral squamous cell carcinomas from Sudanese snuff dippers: lack of association with human papillomavirus infection

In stratified squamous epithelia, altered expression of keratins (Ks) is one possible marker of malignant potential. In the epithelium of the uterine cervix, presence of human papillomaviruses (HPVs) is increasingly regarded as a marker of risk for cervical cancer. However, a similar role in oral cancer and precancer remains controversial. To address these questions, formalin-fixed, paraffin-embedded oral carcinomas from Sudanese snuff dippers (n=14) and oral carcinomas from Sudanese (n=14), Swedish (n=19) and Norwegian (n=41) non-snuff dippers were examined by immunohistochemistry for expression of K types 13, 14 and 19 using monoclonal antibodies. HPV infection was searched for in all the carcinomas by in situ hybridization (ISH) using the cocktail HPV OmniProbe and the ViraType probe. Carcinomas from Sudanese (snuff dippers/non-snuff dippers) were also examined for HPV infection by polymerase chain reaction (PCR) using the general HPV primers GP5+/GP6+. For the oral carcinomas from snuff dippers, moderate to intense expression of K13 (71%; 10/14), K14 (86%; 12/14) and K19 (93%; 13/14) was found. For the oral carcinomas from non-snuff dippers, weak to moderate expression of K13 (64%; 47/74), K14 (43%; 32/74) and K19 (45%; 33/74) was found. HPV DNA was not detected in any of the carcinomas from three countries when examined by ISH. The Sudanese (from snuff dippers/non-snuff dippers) oral carcinomas were also negative for HPV DNA with the PCR. The present study shows that (i) there is a high level of expression of K13, K14 and K19 in oral carcinomas from snuff dippers compared to those from non-snuff dippers, (ii) this high level of expression may arise from dysregulation of keratinocyte proliferation and maturation caused by damaging effects of snuff, (iii) the HPV genome is not found in Sudanese (snuff dippers/non-snuff dippers), Swedish or Norwegian oral carcinomas, and (iv) this may suggest that these viruses do not play a prominent role in the aetiology of oral carcinomas from these countries.     

Spectral discrimination of healthy and Ganoderma-infected oil palms from hyperspectral data

Although hyperspectral remote sensing has been used to study many agricultural phenomena such as crop stress and diseases, the potential use of this technique for detecting Ganoderma disease infestations and damage to oil palms under field conditions has not been explored to date. This research was conducted to investigate the feasibility of using a portable hyperspectral remote-sensing instrument to identify spectral differences between oil-palm leaves with and without Ganoderma infections. Reflectance spectra of samples representative of three classes of disease severity were collected. The most significant bands for spectral discrimination were selected from reflectance spectra and first derivatives of reflectance spectra. The significant wavelengths were identified using one-way analysis of variance. Then, a Jeffries–Matusita (JM) distance measurement was used to determine spectral separability between the classes. A maximum likelihood classifier method was used to classify the three classes based on the most significant wavelength spectral responses, and an error matrix was finally used to assess the accuracy of the classification.     

Stemming Hausa text: using affix-stripping rules and reference look-up

Stemming is a process of reducing a derivational or inflectional word to its root or stem by stripping all its affixes. It is been used in applications such as information retrieval, machine translation, and text summarization, as their pre-processing step to increase efficiency. Currently, there are a few stemming algorithms which have been developed for languages such as English, Arabic, Turkish, Malay and Amharic. Unfortunately, no algorithm has been used to stem text in Hausa, a Chadic language spoken in West Africa. To address this need, we propose stemming Hausa text using affix-stripping rules and reference lookup. We stemmed Hausa text, using 78 affix stripping rules applied in 4 steps and a reference look-up consisting of 1500 Hausa root words. The over-stemming index, under-stemming index, stemmer weight, word stemmed factor, correctly stemmed words factor and average words conflation factor were calculated to determine the effect of reference look-up on the strength and accuracy of the stemmer. It was observed that reference look-up aided in reducing both over-stemming and under-stemming errors, increased accuracy and has a tendency to reduce the strength of an affix stripping stemmer. The rationality behind the approach used is discussed and directions for future research are identified.     

An artificial neural network-based mathematical model for the prediction of blast-induced ground vibrations

ABSTRACT

This paper presents an artificial neural network (ANN) based mathematical model for the prediction of blast-induced ground vibrations using the data obtained from the literature. A feed-forward back-propagation multi-layer perceptron (MLP) was adopted, and the Levenberg–Marquardt algorithm was used in training the network. The powder factor, the maximum charge per delay, and distance from blasting face to monitoring point are the input variables. The peak particle velocity (PPV) is the targeted output variable. The model was then formulated using the weights and biases output from the ANN simulation. Multilinear regression (MLR) analysis was also performed using the same number of datasets, as in the case of ANN. The quality of the proposed ANN-based model was also tested with another 14 datasets outside the one used in developing the models and compared with more classical models. The coefficient of the determination (R²) of the proposed ANN-based model was the highest.     

Polycarbonate/silica nanocomposite membranes: Fabrication,characterization, and performance evaluation

Polycarbonate/silica nanocomposite membranes at low silica loading were fabricated by solution blending and solvent evaporation technique. The functionalized silica nanoparticles used were synthesized by co‐condensing hydrolyzed tetraethylorthosilicate with 3‐aminopropyl trimethoxysilane in the sol–gel process. The membranes morphology, composition, surface, structure, thermal and mechanical properties were analyzed by the standard characterization techniques. The gas permeation tests were conducted in four‐channel permeation cells. Field emission scanning electron microscopy results reveal that membranes above 3 wt % silica content formed distinguishable voids and agglomerates. Fair distribution of silica nanoparticles and absence of residual solvents were observed by energy dispersive X‐ray and thermogravimetric analysis. Fourier transform infrared spectroscopy spectra confirmed the presence of new functional groups (N? H) and (O? H) bonds. The X‐ray diffraction pattern revealed the polymer‐particle interactions, the formation of rigidified polymer chain, and nanostructured silicon crystals. Further, the thermogravimetric analysis results revealed thermal stability enhancement while differential scanning calorimetry results of increased glass transition temperatures confirmed the presence of rigidified polymer chain. Furthermore, enhancements in mechanical strength of the membranes were observed. Moreover, at all feed pressures, increased CO₂, N₂, and CH₄ gas permeation was observed. At 6 bar feed pressure, the CO₂/N₂ and CO₂/CH₄ ideal selectivities of PC membranes with 3 wt % silica loading have increased from 19.2 to 38.0 and 29.2, respectively. © 2017 Wiley Periodicals, Inc. J. Appl. Polym. Sci. 2017 , 134 , 45310.