Turnover Prediction in a Call Center: Behavioral Evidence of Loss Aversion using Random Forest and Naïve Bayes Algorithms

It is well known that call centers suffer from high levels of employee turnover; however, call centers are services that have excellent operational records of telemarketing activities performed by each employee. With this information, we propose to use the Random Forest and the naïve Bayes algorithms to build classifiers and predict turnover of the sales agents. The results of 2407 sales agents’ operational performance records showed that, although the naïve Bayes is much simpler than Random Forest, both classifiers performed similarly, achieving interesting accuracy rates in turnover prediction. Moreover, evidence was found that incorporating performance differences over time increases significantly the accuracy of the predictive models up to 85%, with the naïve Bayes being quite competitive with the Random Forest classifier when the amount of information is increased. The results obtained in this study could be useful for management decision-making to monitor and identify potential turnover due to poor performance, and therefore, to take a preventive action.     

Job performance prediction in a call center using a naive Bayes classifier

This study presents an approach to predict the performance of sales agents of a call center dedicated exclusively to sales and telemarketing activities. This approach is based on a naive Bayesian classifier. The objective is to know what levels of the attributes are indicative of individuals who perform well. A sample of 1037 sales agents was taken during the period between March and September of 2009 on campaigns related to insurance sales and service pre-paid phone services, to build the naive Bayes network. It has been shown that, socio-demographic attributes are not suitable for predicting performance. Alternatively, operational records were used to predict production of sales agents, achieving satisfactory results. In this case, the classifier training and testing is done through a stratified tenfold cross-validation. It classified the instances correctly 80.60% of times, with the proportion of false positives of 18.1% for class no (does not achieve minimum) and 20.8% for the class yes (achieves equal or above minimum acceptable). These results suggest that socio-demographic attributes has no predictive power on performance, while the operational information of the activities of the sale agent can predict the future performance of the agent.     

零售商销售代理的协同激励对供应链绩效影响

以制造商、零售商和销售代理组成的三级供应链系统为背景,分析销售代理的协同激励对供应链绩效的影响.应用委托代理理论,针对协同激励和传统激励两种情形设计相应的契约,并作对比分析.研究结果表明,协同激励时,制造商收益增加,零售商收益减少,整个供应链系统收益增加.同时可以发现,制造商给予零售商的分成系数和零售商给予销售代理的佣金均减少,底薪的变化则由销售代理的风险规避系数而定.     

Quantifying worker interactions and characterizing worker performance through ubiquitous IoT

Cloud technologies and artificial intelligence are transforming call centers into intelligent relationship hubs. Such transformation requires call-center employees to be strategically connected based on the distribution of expertise. However, today’s organizations lack a real-time method to agilely and pervasively map knowledge-distribution and optimize information flow. We present a pilot study showing that the interactions captured by Zigbee and infrared (IR) Internet of things (IoT) sensors on sociometric badges (a business-card-sized printed circuit board) worn by 36 employees in a call center can be used to characterize interactions and their impact on employee performance. Specifically, we quantify an employee’s centrality, weak ties, and strong ties from sensor network data and analyze the effects on average task-processing time. Further analysis reveals insights into interactions among workers that were previously limited by coarse qualitative data in survey studies. This study points to the potential of a “living lab” approach for investigating the effects of employee interactions and behaviors on their performance quantitatively in the real word using ubiquitous IoT.     

LEARNING DRIFTING NEGOTIATIONS

In this work, we propose the use of drift detection techniques for learning offer policies in multiissue, bilateral negotiation. Several works aiming to develop adaptive trading agents have been proposed. Such agents are capable of learning their competitors' utility values and functions, thereby obtaining better results in negotiation. However, the learning mechanisms generally used disregard possible changes in a competitor's offer/counter-offer policy. In that case, the agent's performance may decrease drastically. The agent then needs to restart the learning process, as the model previously learned is no longer valid. Drift detection techniques can be used to detect changes in the current offers model and quickly update it. In this work, we demonstrate with simulated data that drift detection algorithms can be used to build adaptive trading agents and offer a number of advantages over the techniques mostly used in this problem. The results obtained with the algorithm IB3 (instance-based) show that the agent's performance can be rapidly recovered even when changes interesting to the competitor are abrupt, moderate, or gradual.     

Performability evaluation of emergency call center

Emergency call centers serve people in utmost circumstances; hence they should be highly dependable. Availability and performance are key aspects in call centers. Architectural models, service policies, and redundancies are attributes for defining and evaluating the overall operation of such systems. In this work, we performed the performability evaluation of an important emergency call center located in a large city in Brazil. The results showed that the system’s downtime is considered high, so, we proposed a new architecture to improve the system’s performability. The results obtained in this paper can be used to provide support for decisions on interventions in the emergency call center to improve its performability. It is expected that the models presented in this paper will be useful in a variety of emergency call centers.     

External environment of an autonomous manufacturing agent: dynamics and representation

A formal representation of an agent's external environment is proposed, that can be used in information flow analysis. An autoregressive process of second order is applied as the mathematical tool. The dependency X(t) = f[X[0), X(1)] is derived and the characteristic features of the dynamics of the external environment are discussed. The proposed formal representation increases the set of tools for modelling and simulation of autonomous functioning manufacturing agents. It allows for the agent rigorous information input definition and assists systems analysts in the evaluation of an information structure using different attributes of the external environment.     

Distributed scheduling to support a call center: A cooperative multiagent approach

This article describes a multiagent system architecture to increase the value of 24-hour-a day call center service. This system supports call centers in making appointments with clients on the basis ofknowledge ofemployees and their schedules. Relevant activities are scheduled for employees in preparation ofsuch appointments. The multiagent system architecture is based on principled design, using the compositional development method for multiagent systems DESIRE. To schedule procedures in which more than one employee is involved, each employee is represented by its own personal assistant agent, and a work manager agent coordinates the schedules of the personal assistant agents and clients through the call center. The multiagent system architecture has been applied to the banking domain, in cooperation with and partially funded by the Rabobank.     

Limited benefit of joint estimation in multi‐agent iterative learning

This paper studies iterative learning control (ILC) in a multi‐agent framework, wherein a group of agents simultaneously and repeatedly perform the same task. Assuming similarity between the agents, we investigate whether exchanging information between the agents improves an individual's learning performance. That is, does an individual agent benefit from the experience of the other agents? We consider the multi‐agent iterative learning problem as a two‐step process of: first, estimating the repetitive disturbance of each agent; and second, correcting for it. We present a comparison of an agent's disturbance estimate in the case of (I) independent estimation, where each agent has access only to its own measurement, and (II) joint estimation, where information of all agents is globally accessible. When the agents are identical and noise comes from measurement only, joint estimation yields a noticeable improvement in performance. However, when process noise is encountered or when the agents have an individual disturbance component, the benefit of joint estimation is negligible. Copyright © 2011 John Wiley and Sons Asia Pte Ltd and Chinese Automatic Control Society     

Attractive Agents Are More Persuasive

Advances in more ubiquitous human–computer interface technologies have resulted in the increased use of virtual agents, thus highlighting the need to understand the impact these artificial entities have on the human user in terms of how the user both perceives and behaves toward these characters. We report the results of an experiment based on the Desert Survival task, where the effect of agent attractiveness on persuasion was tested. The results suggest that the attractive agent was significantly more persuasive in changing the participant's opinion than was the unattractive agent. Furthermore, greater behavioral change was observed when the participant interacted with the attractive agent, in terms of persuasiveness, including the attribution of more positive personality traits toward this agent (the attractiveness stereotype). These findings provide further evidence of the powerful influence of the agent's physical appearance on the human user in an interaction-based context. We conclude by pointing toward future research and potential within the area.     

How Does the Agents' Appearance Affect Users' Interpretation of the Agents' Attitudes: Experimental Investigation on Expressing the Same Artificial Sounds From Agents With Different Appearances

An experimental investigation into how the appearance of an agent such as a robot or PC affects people's interpretations of the agent's attitudes is presented. In general, people are said to create stereotypical agent behavioral models in their minds based on the agents' appearances, and these appearances significantly affect their way of interaction. Therefore, it is quite important to address with the following research question: How does an agent's appearance affect its interactions with people? Specifically, a preliminary experiment was conducted to select eight artificial sounds for which people can estimate two specific primitive attitudes (e.g., positive or negative). Then an experiment was conducted where the participants were presented with the selected artificial sounds through three kinds of agents: a MindStorms robot, AIBO robot, and laptop PC. In particular, the participants were asked to select the correct attitudes based on the sounds expressed by these three agents. The results showed that the participants had better interpretation rates when a PC presented the sounds and lower rates when the MindStorms and AIBO robots presented the sounds, even though the sounds expressed by these agents were the same. The results of this study contribute to the design policy of the interactive agents, such as, What types of appearances should agents have to effectively interact with people, and which kinds of information should these agents express to people?     

A framework for balanced service and cross-selling by using queuing science

Contact centers are complex call centers to handle large volume of inbound, outbound or both types of calls depending on the business purpose. Call centers assume the role of the primary contact medium for many companies from a wide range of industries with their customers or clients. Despite of being seen traditionally as adding cost to the companies’ bottom lines, call centers are now viewed by many companies to turn a service request into an opportunity to sell additional products and services. This sales attempt is called cross-selling. The opportunity to generate profit from an existing customer-base is a key factor for a successful call center. This paper introduces a framework for balancing cross-selling and service activities in a call center setting from a queuing science point of view. The main goal of this study is to introduce a framework to maximize a call center’s performance without degrading the service quality. Our framework is based on the usage of real-time queue characteristics, customer profile information and server-skill set information from a cross-sell point of view.     

移动社交网络中细粒度朋友发现隐私保护机制

在移动社交网络中,用户可以通过匹配彼此的特征属性进行朋友发现,针对单属性管理中心用户属性密钥更容易被攻击者窃取和服务高峰出现的性能瓶颈问题,提出一种由多个属性管理中心、分级管理用户属性子密钥方案.在该方案中,多个属性中心细粒度地管理用户的不同特征属性,并根据用户特征属性生成属性子密钥,交友请求者只有满足交友发起者设置的交友访问策略,才能正确地将各子密钥组合成完整的解密密钥,进而解密存储在交友中心的用户加密数据文件.通过对属性子密钥进行分级分类管理,不仅避免了单属性管理中心容易被攻击而造成的密钥泄漏以及单点故障风险,而且多属性中心协同工作提高了交友匹配计算效率.通过验证方案是否可挑战明文攻击,证明可达到CPA安全,可以有效地保护用户的隐私不被泄露.同时与既有方案进行了充分的对比实验,确保该方案计算开销最小,可以提供良好的用户体验.     

Employee turnover: a neural network solution

In today's working environment, a company's human resources are truly the only sustainable competitive advantage. Product innovations can be duplicated, but the synergy of a company's workforce cannot be replicated. It is for this reason that not only attracting talented employees but also retaining them is imperative for success. The study of employee turnover has attempted to explain why employees leave and how to prevent the drain of employee talent. This paper focuses on using a neural network (NN) to predict turnover. If turnover can be found to be predictable the identification of at-risk employees will allow us to focus on their specific needs or concerns in order to retain them in the workforce. Also, by using a Modified Genetic Algorithm to train the NN we can also identify relevant predictors or inputs, which can give us information about how we can improve the work environment as a whole. This research found that a NNSOA trained NN in a 10-fold cross validation experimental design can predict with a high degree of accuracy the turnover rate for a small mid-west manufacturing company.     

Dimensions of Employee Satisfaction as Determinants of Organizational Commitment in the Hotel Industry

In today's dynamic environment, the role of employees and their level of organizational commitment are becoming increasingly important for business success, which is especially relevant for the service industry in general and for the tourism sector in particular because of problems associated with employee turnover, loyalty strategies with the customer, and so forth. Employee satisfaction is considered an essential ingredient for developing organizational commitment. The main purpose of this study is to analyze the influence of different facets of employee job satisfaction, that is, job conditions, reward system, relations with superiors and co‐workers, organizational human resources (HR) policies, on employees’ organizational commitment in the hotel industry. The research hypotheses are tested with a sample of 760 hotel employees on Gran Canaria, Spain. The results show the relevance of the working conditions, direct rewards, relationships with managers, and the HR policies in order to develop a high level of organizational commitment.     

Separating the wheat from the chaff: An intelligent sales recruitment and benchmarking system

The need for computer-based intelligent techniques for recruitment and retention of employees in a highly competitive global market has grown significantly in the last decade. Salesperson recruitment is a critical task for most organisations. Existing approaches for salesperson recruitment primarily rely on filtering of applications based on selection criteria followed by interviews. Some organisations also include personality testing based on psychometric techniques. The high turnover of salesperson in the industry suggests limited success of these procedures. Additionally, existing approaches lack benchmarking methods. In this paper we describe design and development of an intelligent sales recruitment and benchmarking system (ISRBS) for recruitment and benchmarking of salespersons. ISRBS design represents operation of the findings and outcomes based on actual field studies and random surveys of salespersons as well as development of models for measuring independent and dependent variables related to selling behaviour. The main contributions of the paper are (i) Developing an on line selling behaviour profiling technique based on integration of intelligent system techniques like expert systems and fuzzy sets, psychology based selling behaviour model, and AHP techniques, and (ii) an objective and novel selling behaviour benchmarking technique to facilitate modelling of organisation based benchmarks and cultural fits. An earlier version of this system has been commercially used in the industry in Australia. ISRBS integrates psychology based selling behaviour model with artificial intelligence techniques and soft computing methods for selling behaviour profiling and benchmarking.     

Modelling adaptive multi-agent manufacturing control with discrete event system formalism

This paper introduces an approach for modelling and designing multi-agent control architectures for agile manufacturing using a generic formalism based on a system-theoretic discrete event approach. To describe the details of the modelling strategy, we apply the proposed approach to a multi-agent network for Job flow control in a manufacturing plant. Two interacting types of autonomous controllers, Part Agents and Machine Agents, are in charge of controlling the part flow and the machine processing sequences. Both type of agents are first modelled as atomic discrete event systems and subsequently integrated in the model of the entire network of autonomous controllers. To improve the performance of the network of agents, we introduce a mechanism based on evolutionary algorithms adapting the agents’ decision laws that are encapsulated in agents’ states. Through network simulation, the algorithm continuously searches for effective decision laws, consequently adapting agent's behaviour to the current operational conditions of the manufacturing floor. Simulation results show the potentialities of the approach.     

Are some human ecosystems self-defeating?

Complex patterns of human behaviour are difficult to capture in agent-based simulations of socio-ecological systems. Even knowing each individual agent's strategy at one point in time may not help when trying to predict the collective behaviour of certain systems – e.g. if it is in each agent's best interest to do the opposite of most other agents. In self-defeating situations like these, the collective population of agents may exhibit a panorama of simple or complex behaviour, depending on the extent to which useful information is shared. An extreme example is the bar problem, in which a simulated population of bar attendees oscillates in a seemingly random manner around a critical congestion level. This paper suggests that several resource management problems involving human interactions with ecosystems may possess a self-defeating character. This poses new challenges for integrated resources management. A case in point is the potential over-fishing of fisheries, which is addressed in the paper and likened to a minority game. It is concluded that a mix of innovative and imitative behaviour may be the key to overcoming self-defeating tendencies.     

Personalizing user–agent interaction

Interface agents are computer programs that provide personalized assistance to users with their computer-based tasks. The interface agents developed so far have focused their attention on learning a user's preferences in a given application domain and on assisting him according to them. However, in order to personalize the interaction with users, interface agents should also learn how to best interact with each user and how to provide them assistance of the right sort at the right time. To fulfil this goal, an interface agent has to discover when the user wants a suggestion to solve a problem or deal with a given situation, when he requires only a warning about it and when he does not need any assistance at all. In this work, we propose a learning algorithm, named WoS, to tackle this problem. Our algorithm is based on the observation of a user's actions and on a user's reactions to the agent's assistance actions. The WoS algorithm enables an interface agent to adapt its behavior and its interaction with a user to the user's assistance requirements in each particular context.     

Retracted: The Relationship between Personality Traits and Sales Force Automation Usage: A Preliminary Study

Sales force automation (SFA) came to cognition and officially gained momentum in the 1990s (Taylor, 1993). SFA provides salespeople with speedy information access, diminishing the time needed to prepare for calls and cutting down on the number of follow‐ups, particularly for urgent need of further information support (Taylor, 1993). This study attempts to look specifically at the “utilization–individual impact” link. In particular, the close link of the personality type of the sales representative to the individual performance usage of the sales forces automation systems in the pharmaceutical industry in Taiwan. This study provides insight into how the personality traits of the individual medical representatives of pharmaceutical companies in Taiwan could influence the adoption of innovative technology, the sales force automation system. © 2012 Wiley Periodicals, Inc.