Profit uplift modeling for direct marketing campaigns: approaches and applications for online shops

In order to select “best” customers for a direct marketing campaign, response models are widespread: a sample of customers receives an ad, a catalog, a sample pack, or a discount offer on a test basis. Then, their responses (e.g., website visits, conversions, or revenues) are used to build a predictive model. Finally, this model is applied to all customers in order to select “best” ones for the campaign. However, up to now, only models that reflect website visits, conversions, or revenues have been proposed. In this paper, we discuss the shortcomings of these traditional approaches and propose profit uplift modeling appoaches based on one-stage ordinary regression and random forests as well as two-stage Heckman sample selection and zero-inflated negative binomial regression for parameter estimation. The new approaches demonstrate superiority to the traditional ones when applied to real-world datasets. One dataset reflects recent discount offers of a large online fashion retailer. The other is the well-known Hillstrom dataset that describes two Email campaigns.

Forecasting of Kharif Cereal Production in Odisha by Using Spline Regression Technique

Cereals are prime determinant of agricultural status of the state mainly during kharif season. Forecasting of the production of kharif cereals is of utmost importance to formulate the agricultural policy and strategy of the state. The ARIMA model can be reliably used to forecast for short future periods because uncertainty in prediction increases when done for longer future periods. The predictions obtained from the ordinary regression model are valid only when the relationship between the independent variables and the dependent variable does not change significantly in the future period which can be rarely assumed. It is expected that the spline regression will overcome the respective discrepancies in both ARIMA and ordinary regression techniques of forecasting with the assumption that the future period which needs forecasting follows the same pattern as the last partitioned period. The entire period of data is split into different periods based on the scatter plot of the data The suitable regression models, such as, linear, compound, logarithmic and power model are fitted to the data on area and yield of kharif cereals by using the training set data. Selection of best fit model is done on the basis of overall significance of the model, model diagnostic test for error assumptions and model fit statistics. The selected best fit model is then cross validated with the testing set data. After successful cross validation of the selected best fit models, they are used for forecasting of the future values for their respective variables. The models found to be best fit and thus selected for cross validation purpose are compound spline model for both area and yield of kharif cereals respectively. Forecasting of area, yield and hence production of kharif cereals for six years ahead i.e., for the year 2020-21 to 2025-26 by using the selected best fit model after successful cross validation. The forecast values for production of kharif cereals are found to decrease despite increase in forecast values of yield which is due to decrease in forecast value of area.

Formulating Intuitionistic Fuzzy Regression Models: A Mathematical Programming Approach

This study develops a mathematical programming approach to establish intuitionistic fuzzy regression models (IFRMs) by considering the randomness and fuzziness of intuitionistic fuzzy observations. In contrast to existing approaches, the IFRMs are established in terms of five ordinary regression models representing the components of the estimated triangular intuitionistic fuzzy response variable. The optimal parameters of the five ordinary regression models are determined by solving the proposed mathematical programming problem, which is linearized to make the resolution process efficient. Based on the concepts of randomness and fuzziness in the formulation processes, the proposed approach can improve on existing approaches’ weaknesses with establishing IFRMs, such as the limitation of symmetrical triangular membership (or non-membership) functions, the determination of parameter signs in the model, and the wide spread of the estimated responses. In addition, some numerical explanatory variables included in the intuitionistic fuzzy observations are also allowed in the proposed approach, even though it was developed for intuitionistic fuzzy observations. In contrast to existing approaches, the proposed approach is general and flexible in applications. Comparisons show that the proposed approach outperforms existing approaches in terms of similarity and distance measures.

PLS REGRESSION TO COMPARE THE RELATIVE IMPORTANCE OF WATER QUALITY VARIABLES IN FISH GROWTH

Partial least squares (PLS) is a multivariate dimension reduction technique which is not based on ordinary regression assumptions. The use of PLS regression in life sciences is still a novel concept despite many scientific applications. This paper analyses the influence of physicochemical in the two fish species, Oreochromis jipe and Oreochromis niloticus to determine the cause for their growth difference in the same culture environment. The graphical display of the multi-parameter analysis was performed using a suite of open access R-software packages. The modeling hypothesis was assessed using experimental data collected for the period of 84 days. The findings revealed that significant linear relationship exists between water quality and mean weight of both O. jipe and O.niloticus fish species. Being a crucial study meant to provide baseline information to asses the aquaculture potential O. jipe, we recommend a further study to be conducted on several other predictor variables that can be measured under controlled aquaculture conditions.

Learner performance in the 2009 to 2014 final Grade 12 mathematics examination: A quantile regression approach

The South African education system bears evidence of fluctuations in the final Grade 12 mathematics marks occurring across different learner profiles. This study reflected on the National Senior Certificate (NSC) mathematics results from the Western Cape Education Department for the years 2009 to 2014, the period just after the introduction of the NSC in 2008 and including the updated NSC introduced in 2014. Accordingly, this study aimed to examine the learners’ performance by socio-economic school quintile and education district for the period of 2009 to 2014, for learners in the Western Cape. Instead of the ordinary regression model, we adopted the quantile regression approach to examine the effect of school (national) quintile (NQ) type and education district at different quantiles of learner performance in the mathematics examination. The results showed that there is a significant school quintile type and education district effect on learner performance in NSC mathematics examinations for learners in the Western Cape. In some years, there were no significant performance differences between learners from NQ2 and NQ4 schools in the different quantiles. Similarly, learner performance differences for NQ3 and NQ4 schools were not significant. As we moved from 2009 to 2014, the performance difference between the lower school quintiles and the upper school quintiles narrowed, although the performance differences remained significant. These differences were smallest in 2013. This is a good sign, as it indicates that government efforts and policies, designed to narrow the historical social disparities manifested in the schools, have been somewhat successful. The identification and scrutinising of school quintile type and education district where the gap is wider will assist the government to review policies and interventions to accelerate the transformation.

Critical Examination of the Parametric Approaches to Analysis of the Non-Verbal Human Behavior: A Case Study in Facial Pre-Touch Interaction

A prevailing assumption in many behavioral studies is the underlying normal distribution of the data under investigation. In this regard, although it appears plausible to presume a certain degree of similarity among individuals, this presumption does not necessarily warrant such simplifying assumptions as average or normally distributed human behavioral responses. In the present study, we examine the extent of such assumptions by considering the case of human–human touch interaction in which individuals signal their face area pre-touch distance boundaries. We then use these pre-touch distances along with their respective azimuth and elevation angles around the face area and perform three types of regression-based analyses to estimate a generalized facial pre-touch distance boundary. First, we use a Gaussian processes regression to evaluate whether assumption of normal distribution in participants’ reactions warrants a reliable estimate of this boundary. Second, we apply a support vector regression (SVR) to determine whether estimating this space by minimizing the orthogonal distance between participants’ pre-touch data and its corresponding pre-touch boundary can yield a better result. Third, we use ordinary regression to validate the utility of a non-parametric regressor with a simple regularization criterion in estimating such a pre-touch space. In addition, we compare these models with the scenarios in which a fixed boundary distance (i.e., a spherical boundary) is adopted. We show that within the context of facial pre-touch interaction, normal distribution does not capture the variability that is exhibited by human subjects during such non-verbal interaction. We also provide evidence that such interactions can be more adequately estimated by considering the individuals’ variable behavior and preferences through such estimation strategies as ordinary regression that solely relies on the distribution of their observed behavior which may not necessarily follow a parametric distribution.

Analysis of Preferable Occupation between Merchant and Islamic Bank Clerk in Bank Syariah Mandiri (BSM) Based on a Hadith about Riba

This paper investigates the relationship between occupation to interest rate between Islamic bank clerk and merchant who export or import goods. As Prophet Muhammad ever said that at the end people still eat the “the dust of riba” event they want to avoid it, this implies that what people’s occupation will also be affected by the rate of interest rate. Introducing two model that represent the relationship between two occupation, merchant and Islamic bank clerks, and using ordinary regression, this study finds that statistically, as people want to be a merchant, they will gain much benefit when the rate of interest raise, however, when people prefer to become Islamic bank clerk (especially in BSM as case study), their income will harm as the rate of interest increase. In addition, when the interest rate rise, the benefit which is gained by merchant much greater than the loss of Islamic bank profit (BSM) because of the higher coefficient value of merchant rather than Islamic bank profit (BSM).

Latent variable models for multi-species counts modeling in ecology

Herliansyah R, Fitria I. 2018. Latent variable models for multi-species counts modeling in ecology. Biodiversitas 19: 1871-1876. High-dimensional multi-species counts are often collected in ecology to understand the spatial distribution over different locations and to study effects of environmental changes. Modeling multivariate abundance is challenging as we need to consider the possibility of interactions across species. Latent variable models are the recent popular approaches in statistical ecology to address such issue that has a similar framework to ordinary regression models. In this paper, we employed the poisson distribution for modeling count responses and a negative binomial distribution for more frequent zeros in observations. The implementation of a latent variable model, Generalized Linear Latent Variable Models (GLLVMs), was demonstrated on multi-species counts of endemic bird species collected in 37 different sites in Central Kalimantan, Indonesia. The main objectives were to study the effect of logging activities on abundance of endemic species and their interactions and to observe the habitat preference of certain species. Our study found that out of four endemic species, Alophoixus bres and Eurylaimus javanicus species were significantly affected by logging activities. The sign of parameters was negative indicating the logging activities in 1989 and 1993 bring significantly negative impacts on those species. The interaction created among species was strongly negative for major endemic species especially Alophoixus bres and Eurylaimus javanicus that prefer living in primary forest than in logging areas.

A Bayesian Quantile Modeling for Spatiotemporal Relative Risk: An Application to Adverse Risk Detection of Respiratory Diseases in South Carolina, USA

Quantile modeling has been seen as an alternative and useful complement to ordinary regression mainly focusing on the mean. To directly apply quantile modeling to areal data the discrete conditional quantile function of the data can be an issue. Although jittering by adding a small number from a uniform distribution to impose pseudo-continuity has been proposed, the approach can have a great influence on responses with small values. Thus we proposed an alternative to model the quantiles of relative risk for spatiotemporal areal health data within a Bayesian framework using the log-Laplace distribution. A simulation study was conducted to assess the performance of the proposed method and examine whether the model could robustly estimate quantiles of spatiotemporal count data. To perform a test with a real data example, we evaluated the potential application of clustering under the proposed log-Laplace and mean regression. The data were obtained from the total number of emergency room discharges for respiratory conditions, both infectious and non-infectious diseases, in the U.S. state of South Carolina in 2009. From both simulation and case studies, the proposed quantile modeling demonstrated potential for broad applicability in various areas of spatial health studies including anomaly detection.

THE INFLATED VALUATION PROBLEM IN VALENCIA, SPAIN, AND IMPLICATIONS FOR FIRM SIZE

Home purchase-sale prices have been widely modeled by several authors. Nonetheless, other values exist, such as home mortgage appraisal values, used by financial institutions, which have played a key role in the recent financial crisis. This article attempts to model the appraisal price of one m2 of residential properties obtained by 31 appraisal companies in Valencia (Spain). Mortgage appraisal values of 17 007 residential properties were used for this purpose. Spatial autocorrelation was detected in both the data and residuals of the ordinary regression model, which justified using spatial regression models. Of the four employed models, the error model offered the best results. Significant differences were found among appraisal companies, which varied as much as 83% for some. Generally speaking, small appraisal companies obtained higher over-valuation percentages, which confirms their situation of weakness. The fact that over-valuations exist in mortgage securities is a high risk for a stable financial system.