Chaochang Chiu | Yuan Ze University (original) (raw)

Papers by Chaochang Chiu

Research paper thumbnail of Application of network DEA in the assessment of operating performances of farmers associations in Taiwan

Journal of Interdisciplinary Mathematics, 2012

Abstract This paper applies Network Data Envelopment Analysis (Network DEA) in the assessment of ... more Abstract This paper applies Network Data Envelopment Analysis (Network DEA) in the assessment of the operating performances of 199 farmers’ associations in Taiwan during the period 2006˜2008. The assessment concerns four departments: marketing & distributions, financing (credit), insurance, and agricultural promotion; it also refers to the Malmquist Productivity Index (MPI) to measure the productivity changes of different associations over time and performs variance analyses to highlight the strengths and weaknesses of individ-ual farmers’ associations. The purposes include providing a reference for their operational strategies and setting an objective benchmark for operational efficiency. This paper finds that as far as MPI is concerned, (2) village-type farmers associations outperformed city-type and hybrid-type farmers associations in 2006˜2008. (2) In 2006˜2008, only farmers associations in central Taiwan saw their productivity improve. According to Network DEA, (1) the total Efficiency of city-type farmers associations is 0.553. The efficiency of hybrid-type farmers associations is lower than the average. Village-type farmers associations report a better-than-average efficiency in marketing & distribution and insurance functions. (2) Geographically speaking, farmers associations in northern Taiwan perform the best, followed by those in central Taiwan and then those in southern Taiwan.

Research paper thumbnail of A Convolutional Neural Networks Approach with Infrared Array Sensor for Bed-Exit Detection

Among various kinds of falling prevention measures, bed exit alarm mechanism has raised serious a... more Among various kinds of falling prevention measures, bed exit alarm mechanism has raised serious attention recently. In particular, the recent inflow of innovative ICT advancement from Internet of Things, wearable technology, and artificial intelligence have shed on more possibility in realizing effective bed exit alarm systems. This research proposes a deep learning algorithm to construct the bed exit detection model using monitored behavior information collected from the infrared array sensor. Based on the preliminary experiment results, the bed-exit events can be recognized with 92% accuracy, 99% for precision and 97% for recall rate. This approach also has its advantages in low device costs, less data storage needed, less spacial resolution without privacy and legal concerns, and unaffected performance in various lighting conditions.

Research paper thumbnail of A hybrid Latent Dirichlet Allocation approach for topic classification

Many classification techniques can automatically summarize text into topics and accordingly ident... more Many classification techniques can automatically summarize text into topics and accordingly identify topic terms from the online reviews. Among these techniques Latent Dirichlet Allocation (LDA) and Latent Semantic Analysis (LSA) are some of the most often employed approaches. LDA is a probability generated model that projects a document into the topic space using Dirichlet Distribution, and each topic is a collection of words of the probability distribution. As the LDA extracted topics are often implicit, this study first applies LDA to examine the topics of online reviews for game apps in a supervised way. To improve the topic classification performance for LDA, this study proposes a hybrid LDA approach to use Genetic Algorithm (GA) in discovering optimal weights for LDA topics.

Research paper thumbnail of A case-based reasoning system for PCB defect prediction

Expert Systems With Applications, May 1, 2005

The manufacturing process for a new Printed Circuit Board (PCB) design is often instable and migh... more The manufacturing process for a new Printed Circuit Board (PCB) design is often instable and might generate a number of defects during the complicated production process. Defects reduce the yield rate and increase the production costs. Although skilled engineers can predict the possible defect items for a new PCB product, this approach requires strong engineering experience and is time consuming.

Research paper thumbnail of Smart Rabbit – A Wearable Device As An Intelligent Pacer for Marathon Runners

This paper introduces a wearable device that aims to facilitate marathon runners to maintain the ... more This paper introduces a wearable device that aims to facilitate marathon runners to maintain the most decent running pace. We adopt deep learning approach to predict the runner's heart rate. This derived information is combined with existing altitude and road conditions to generate the most suitable pacing suggestion to coach the runner for the best performance. In addition, we have developed a face recognition mechanism to help runners to easily identify and retrieve their photos shot during running in the cloud photo bank with the aid of a personal cell phone instantaneously. Our pilot system has been proven for its feasibility in both the heart rate prediction performance and photo identification rate.

Research paper thumbnail of The Hybrid of PSO and SOM for Blog Success Prediction

Research paper thumbnail of A purchase-based market segmentation methodology

Expert Systems with Applications, 2004

Market segmentation is critical for a good marketing and customer relationship management program... more Market segmentation is critical for a good marketing and customer relationship management program. Traditionally, a marketer segments a market using general variables such as customer demographics and lifestyle. However, several problems have been identified and make the segmentation result unreliable. This paper develops a novel market segmentation methodology based on product specific variables such as purchased items and the associative monetary expenses from the transactional history of customers to resolve these problems. A purchasebased similarity measure, clustering algorithm, and clustering quality function are defined in this paper. A genetic algorithm approach is adopted to ensure that customers in the same cluster have the closest purchase patterns. After completing segmentation, a designated RFM model is used to analyze the relative profitability of each customer cluster. The findings from a practical marketing implementation study will also be discussed.

Research paper thumbnail of Linking Customer Retention to Intelligent Technology: An Optimization Approach

Marketing managers in the telecommunication sectors are confronted with considerable complexity. ... more Marketing managers in the telecommunication sectors are confronted with considerable complexity. They have to make decisions about the optimum combination of products or offerings, customer groups and the means of interacting with potential customers. Further, in saturated markets such as mobile telephony, it is increasingly important to retain customers potentially to churn. On the optimal campaign planning, this research describes how the customer survey was conducted for those potentially churning customers based on which an optimal campaign planning was followed. This research engages with the subjects of customer retention from the perspective of a major mobile operator in Taiwan. Customers’ preferences with C&C (campaign offer and communication channel) were predicted and input for further analysis for target selection optimization. These models was proved novel in an organizational prototype project suggesting that the use of the hybrid of data mining and optimization approac...

Research paper thumbnail of A Proposed Supervised Clustering Approach for the Identification of Concerned HIV-Related Messages in Web Forums

Research paper thumbnail of Text mining self-disclosing health information for public health service

Journal of the Association for Information Science and Technology, 2014

Understanding specific patterns or knowledge of selfdisclosing health information could support p... more Understanding specific patterns or knowledge of selfdisclosing health information could support public health surveillance and healthcare. This study aimed to develop an analytical framework to identify selfdisclosing health information with unusual messages on web forums by leveraging advanced text-mining techniques. To demonstrate the performance of the proposed analytical framework, we conducted an experimental study on 2 major human immunodeficiency virus (HIV)/acquired immune deficiency syndrome (AIDS) forums in Taiwan. The experimental results show that the classification accuracy increased significantly (up to 83.83%) when using features selected by the information gain technique. The results also show the importance of adopting domain-specific features in analyzing unusual messages on web forums. This study has practical implications for the prevention and support of HIV/AIDS healthcare. For example, public health agencies can re-allocate resources and deliver services to people who need help via social media sites. In addition, individuals can also join a social media site to get better suggestions and support from each other.

Research paper thumbnail of A Effectiveness Study of Korea Tourism Advertisement Based on Semiotic Analysis of Roland Barthes for Vietnamese International Students

Peter Drucker Society of Korea, 2016

Research paper thumbnail of A Proposed Supervised Clustering Appproach for the Identification of Concerned HIV-Related Messages in Web Forums

Research paper thumbnail of App Review Analytics Of Free Games Listed On Google Play

Smartphones have become popular in recent years; in turn, the number of application developers an... more Smartphones have become popular in recent years; in turn, the number of application developers and publishers has grown rapidly. To understand users’ app preferences, many platforms such as Google Play provide different mechanism that allows users to rank apps. However, more detailed insights on user’s feelings, experiences, critiques, suggestions, or preferences are missing due to a lack of additional written comments. This research attempts to investigate the review analytics of Android games listed on Google Play using a proposed text analytic approach to extract all user reviews from game apps in Chinese. A total of 207,048 reviews of 4,268 free games from February to March 2013 are extracted and analyzed according to various metrics including game type and game attribute. The findings indicate there is high dependency between users’ gender and game type, males and females have differing opinions on game attributes. In particular, users of different game types prefer different g...

Research paper thumbnail of The System Simulation with Optimization Mechanism for Option Pricing

The spectral energy distribution (SED) sequence for type Ia supernovae (SN Ia) is modeled by an a... more The spectral energy distribution (SED) sequence for type Ia supernovae (SN Ia) is modeled by an artificial neural network. The SN Ia luminosity is characterized as a function of phase, wavelength, a color parameter and a decline rate parameter. After training and testing the neural network, the SED sequence could give both the spectrum with wavelength range from 3000Å to 8000Å and the light curve with phase from 20 days before to 50 days after the maximum luminosity for the supernovae with different colors and decline rates. Therefore, we call this the Artificial Neural Network Spectral Light Curve Template (ANNSLCT) model. We retrain the Joint Light-curve Analysis (JLA) supernova sample by using the ANNSLCT model and obtain the parameters for each supernova to make a constraint on the cosmological ΛCDM model. We find that the best fitting values of these parameters are almost the same as those from the JLA sample trained with the Spectral Adaptive Lightcurve Template 2 (SALT2) model. So we believe that the ANNSLCT model could be used to analyze a large number of SN Ia multi-color light curves measured in the current and future observational projects.

Research paper thumbnail of A Study of Online Customer Loyalty Based on the Theory of Planned Behavior

Research paper thumbnail of Usability Evaluation And Correspondence Analysis Of Smartphone Operating Systems

User-centered smartphone interface design is important so that consumers can easily learn about a... more User-centered smartphone interface design is important so that consumers can easily learn about and begin to use newly purchased smartphones. This study first evaluates the effectiveness, efficiency, and user satisfaction of smartphone interfaces in Taiwan in terms of two representational operating systems: Android and iPhone OS. The usability evaluation includes observational experiments, user questionnaires, and the Wilcoxon sign-rank test. This study then conducts the correspondence analysis to summarize positive/ negative evaluations of usage and specification for smartphone OS. In observational experiments, a total of 48 participants with no previous smartphone experience on using smartphone with Android and iPhone OS are asked to perform five common phone tasks. All experiments are recorded and observed. Results and observations are discussed to attain a closer match between user needs and the performance of smartphone OS.

Research paper thumbnail of The Analysis of Hotel Customer Generated Contents in Weblogs

1 Abstract. This research applies text mining approach to reveal and analyze the customers’ comme... more 1 Abstract. This research applies text mining approach to reveal and analyze the customers’ comments about price, food, facilities, environment, and service of the hotels in Taiwan. We adopt a proposed heuristic n-phrase rule to identify the polarity customer’s opinions and provide position maps to visualize the pros and cons of respective hotels. The research findings about different views and opinions of customers’ perception or experiences with the hotels may provide hotel management with useful insights of customer feedbacks and clues for future improvement.

Research paper thumbnail of Enhancing BEF Luminance for TFT-LCD Industries using the Hybrid Approach of Prediction and Optimization Techniques

A Backlight Module is the main source supply of liquid crystal display components. Its power cons... more A Backlight Module is the main source supply of liquid crystal display components. Its power consumption is about half the liquid crystal display. Among the Backlight Module, BEF / DBEF enhances luminance 50% to 100%. Therefore BEF (Brightness Enhancement Film) / DBEF (Dual Brightness Enhancement Film) becomes the main critical component of Backlight Module. The assurance of sustaining balance for both the high luminance with low consumption is a challenging research issue for industries. Although in the past there many research works have been done about the light guide plate and light source of the Backlight Module, however rare research has been reported about enhancing the brightness film based on constrained production costs. Since there are many manufacturing factors influencing BEF's luminance, in this study we attempt to find out the suitable manufacturing control parameters by employing the hybrid approach of both prediction and optimization techniques. The neural netwo...

Research paper thumbnail of Personal asset allocation using particle swarm optimization approach

Research paper thumbnail of Monthly car sales prediction using Internet Word-of-Mouth (eWOM)

2017 IEEE International Conference on INnovations in Intelligent SysTems and Applications (INISTA), 2017

Internet's mature has changed the behavior of consumers. Most of consumers before purchase wi... more Internet's mature has changed the behavior of consumers. Most of consumers before purchase will queries opinion on the Internet. Vehicle is high priced and durable merchandise, so consumer would be more prudent to view Internet opinion before they buy. Past research has pointed out eWOM (electronic Word-of-Mouth) and customer satisfaction will influence purchasing decision, 70% of consumers believe eWOM. In this study, we utilizes economic indexes, Internet “word of mouth”, and Google Trends variables created by key word searches to forecast the monthly sales volumes of a given model of car. The experiment results demonstrate that the proposed GA/KNN model has the highest predictive power in terms of Mean Absolute Percentage Error (MAPE).

Research paper thumbnail of Application of network DEA in the assessment of operating performances of farmers associations in Taiwan

Journal of Interdisciplinary Mathematics, 2012

Abstract This paper applies Network Data Envelopment Analysis (Network DEA) in the assessment of ... more Abstract This paper applies Network Data Envelopment Analysis (Network DEA) in the assessment of the operating performances of 199 farmers’ associations in Taiwan during the period 2006˜2008. The assessment concerns four departments: marketing & distributions, financing (credit), insurance, and agricultural promotion; it also refers to the Malmquist Productivity Index (MPI) to measure the productivity changes of different associations over time and performs variance analyses to highlight the strengths and weaknesses of individ-ual farmers’ associations. The purposes include providing a reference for their operational strategies and setting an objective benchmark for operational efficiency. This paper finds that as far as MPI is concerned, (2) village-type farmers associations outperformed city-type and hybrid-type farmers associations in 2006˜2008. (2) In 2006˜2008, only farmers associations in central Taiwan saw their productivity improve. According to Network DEA, (1) the total Efficiency of city-type farmers associations is 0.553. The efficiency of hybrid-type farmers associations is lower than the average. Village-type farmers associations report a better-than-average efficiency in marketing & distribution and insurance functions. (2) Geographically speaking, farmers associations in northern Taiwan perform the best, followed by those in central Taiwan and then those in southern Taiwan.

Research paper thumbnail of A Convolutional Neural Networks Approach with Infrared Array Sensor for Bed-Exit Detection

Among various kinds of falling prevention measures, bed exit alarm mechanism has raised serious a... more Among various kinds of falling prevention measures, bed exit alarm mechanism has raised serious attention recently. In particular, the recent inflow of innovative ICT advancement from Internet of Things, wearable technology, and artificial intelligence have shed on more possibility in realizing effective bed exit alarm systems. This research proposes a deep learning algorithm to construct the bed exit detection model using monitored behavior information collected from the infrared array sensor. Based on the preliminary experiment results, the bed-exit events can be recognized with 92% accuracy, 99% for precision and 97% for recall rate. This approach also has its advantages in low device costs, less data storage needed, less spacial resolution without privacy and legal concerns, and unaffected performance in various lighting conditions.

Research paper thumbnail of A hybrid Latent Dirichlet Allocation approach for topic classification

Many classification techniques can automatically summarize text into topics and accordingly ident... more Many classification techniques can automatically summarize text into topics and accordingly identify topic terms from the online reviews. Among these techniques Latent Dirichlet Allocation (LDA) and Latent Semantic Analysis (LSA) are some of the most often employed approaches. LDA is a probability generated model that projects a document into the topic space using Dirichlet Distribution, and each topic is a collection of words of the probability distribution. As the LDA extracted topics are often implicit, this study first applies LDA to examine the topics of online reviews for game apps in a supervised way. To improve the topic classification performance for LDA, this study proposes a hybrid LDA approach to use Genetic Algorithm (GA) in discovering optimal weights for LDA topics.

Research paper thumbnail of A case-based reasoning system for PCB defect prediction

Expert Systems With Applications, May 1, 2005

The manufacturing process for a new Printed Circuit Board (PCB) design is often instable and migh... more The manufacturing process for a new Printed Circuit Board (PCB) design is often instable and might generate a number of defects during the complicated production process. Defects reduce the yield rate and increase the production costs. Although skilled engineers can predict the possible defect items for a new PCB product, this approach requires strong engineering experience and is time consuming.

Research paper thumbnail of Smart Rabbit – A Wearable Device As An Intelligent Pacer for Marathon Runners

This paper introduces a wearable device that aims to facilitate marathon runners to maintain the ... more This paper introduces a wearable device that aims to facilitate marathon runners to maintain the most decent running pace. We adopt deep learning approach to predict the runner's heart rate. This derived information is combined with existing altitude and road conditions to generate the most suitable pacing suggestion to coach the runner for the best performance. In addition, we have developed a face recognition mechanism to help runners to easily identify and retrieve their photos shot during running in the cloud photo bank with the aid of a personal cell phone instantaneously. Our pilot system has been proven for its feasibility in both the heart rate prediction performance and photo identification rate.

Research paper thumbnail of The Hybrid of PSO and SOM for Blog Success Prediction

Research paper thumbnail of A purchase-based market segmentation methodology

Expert Systems with Applications, 2004

Market segmentation is critical for a good marketing and customer relationship management program... more Market segmentation is critical for a good marketing and customer relationship management program. Traditionally, a marketer segments a market using general variables such as customer demographics and lifestyle. However, several problems have been identified and make the segmentation result unreliable. This paper develops a novel market segmentation methodology based on product specific variables such as purchased items and the associative monetary expenses from the transactional history of customers to resolve these problems. A purchasebased similarity measure, clustering algorithm, and clustering quality function are defined in this paper. A genetic algorithm approach is adopted to ensure that customers in the same cluster have the closest purchase patterns. After completing segmentation, a designated RFM model is used to analyze the relative profitability of each customer cluster. The findings from a practical marketing implementation study will also be discussed.

Research paper thumbnail of Linking Customer Retention to Intelligent Technology: An Optimization Approach

Marketing managers in the telecommunication sectors are confronted with considerable complexity. ... more Marketing managers in the telecommunication sectors are confronted with considerable complexity. They have to make decisions about the optimum combination of products or offerings, customer groups and the means of interacting with potential customers. Further, in saturated markets such as mobile telephony, it is increasingly important to retain customers potentially to churn. On the optimal campaign planning, this research describes how the customer survey was conducted for those potentially churning customers based on which an optimal campaign planning was followed. This research engages with the subjects of customer retention from the perspective of a major mobile operator in Taiwan. Customers’ preferences with C&C (campaign offer and communication channel) were predicted and input for further analysis for target selection optimization. These models was proved novel in an organizational prototype project suggesting that the use of the hybrid of data mining and optimization approac...

Research paper thumbnail of A Proposed Supervised Clustering Approach for the Identification of Concerned HIV-Related Messages in Web Forums

Research paper thumbnail of Text mining self-disclosing health information for public health service

Journal of the Association for Information Science and Technology, 2014

Understanding specific patterns or knowledge of selfdisclosing health information could support p... more Understanding specific patterns or knowledge of selfdisclosing health information could support public health surveillance and healthcare. This study aimed to develop an analytical framework to identify selfdisclosing health information with unusual messages on web forums by leveraging advanced text-mining techniques. To demonstrate the performance of the proposed analytical framework, we conducted an experimental study on 2 major human immunodeficiency virus (HIV)/acquired immune deficiency syndrome (AIDS) forums in Taiwan. The experimental results show that the classification accuracy increased significantly (up to 83.83%) when using features selected by the information gain technique. The results also show the importance of adopting domain-specific features in analyzing unusual messages on web forums. This study has practical implications for the prevention and support of HIV/AIDS healthcare. For example, public health agencies can re-allocate resources and deliver services to people who need help via social media sites. In addition, individuals can also join a social media site to get better suggestions and support from each other.

Research paper thumbnail of A Effectiveness Study of Korea Tourism Advertisement Based on Semiotic Analysis of Roland Barthes for Vietnamese International Students

Peter Drucker Society of Korea, 2016

Research paper thumbnail of A Proposed Supervised Clustering Appproach for the Identification of Concerned HIV-Related Messages in Web Forums

Research paper thumbnail of App Review Analytics Of Free Games Listed On Google Play

Smartphones have become popular in recent years; in turn, the number of application developers an... more Smartphones have become popular in recent years; in turn, the number of application developers and publishers has grown rapidly. To understand users’ app preferences, many platforms such as Google Play provide different mechanism that allows users to rank apps. However, more detailed insights on user’s feelings, experiences, critiques, suggestions, or preferences are missing due to a lack of additional written comments. This research attempts to investigate the review analytics of Android games listed on Google Play using a proposed text analytic approach to extract all user reviews from game apps in Chinese. A total of 207,048 reviews of 4,268 free games from February to March 2013 are extracted and analyzed according to various metrics including game type and game attribute. The findings indicate there is high dependency between users’ gender and game type, males and females have differing opinions on game attributes. In particular, users of different game types prefer different g...

Research paper thumbnail of The System Simulation with Optimization Mechanism for Option Pricing

The spectral energy distribution (SED) sequence for type Ia supernovae (SN Ia) is modeled by an a... more The spectral energy distribution (SED) sequence for type Ia supernovae (SN Ia) is modeled by an artificial neural network. The SN Ia luminosity is characterized as a function of phase, wavelength, a color parameter and a decline rate parameter. After training and testing the neural network, the SED sequence could give both the spectrum with wavelength range from 3000Å to 8000Å and the light curve with phase from 20 days before to 50 days after the maximum luminosity for the supernovae with different colors and decline rates. Therefore, we call this the Artificial Neural Network Spectral Light Curve Template (ANNSLCT) model. We retrain the Joint Light-curve Analysis (JLA) supernova sample by using the ANNSLCT model and obtain the parameters for each supernova to make a constraint on the cosmological ΛCDM model. We find that the best fitting values of these parameters are almost the same as those from the JLA sample trained with the Spectral Adaptive Lightcurve Template 2 (SALT2) model. So we believe that the ANNSLCT model could be used to analyze a large number of SN Ia multi-color light curves measured in the current and future observational projects.

Research paper thumbnail of A Study of Online Customer Loyalty Based on the Theory of Planned Behavior

Research paper thumbnail of Usability Evaluation And Correspondence Analysis Of Smartphone Operating Systems

User-centered smartphone interface design is important so that consumers can easily learn about a... more User-centered smartphone interface design is important so that consumers can easily learn about and begin to use newly purchased smartphones. This study first evaluates the effectiveness, efficiency, and user satisfaction of smartphone interfaces in Taiwan in terms of two representational operating systems: Android and iPhone OS. The usability evaluation includes observational experiments, user questionnaires, and the Wilcoxon sign-rank test. This study then conducts the correspondence analysis to summarize positive/ negative evaluations of usage and specification for smartphone OS. In observational experiments, a total of 48 participants with no previous smartphone experience on using smartphone with Android and iPhone OS are asked to perform five common phone tasks. All experiments are recorded and observed. Results and observations are discussed to attain a closer match between user needs and the performance of smartphone OS.

Research paper thumbnail of The Analysis of Hotel Customer Generated Contents in Weblogs

1 Abstract. This research applies text mining approach to reveal and analyze the customers’ comme... more 1 Abstract. This research applies text mining approach to reveal and analyze the customers’ comments about price, food, facilities, environment, and service of the hotels in Taiwan. We adopt a proposed heuristic n-phrase rule to identify the polarity customer’s opinions and provide position maps to visualize the pros and cons of respective hotels. The research findings about different views and opinions of customers’ perception or experiences with the hotels may provide hotel management with useful insights of customer feedbacks and clues for future improvement.

Research paper thumbnail of Enhancing BEF Luminance for TFT-LCD Industries using the Hybrid Approach of Prediction and Optimization Techniques

A Backlight Module is the main source supply of liquid crystal display components. Its power cons... more A Backlight Module is the main source supply of liquid crystal display components. Its power consumption is about half the liquid crystal display. Among the Backlight Module, BEF / DBEF enhances luminance 50% to 100%. Therefore BEF (Brightness Enhancement Film) / DBEF (Dual Brightness Enhancement Film) becomes the main critical component of Backlight Module. The assurance of sustaining balance for both the high luminance with low consumption is a challenging research issue for industries. Although in the past there many research works have been done about the light guide plate and light source of the Backlight Module, however rare research has been reported about enhancing the brightness film based on constrained production costs. Since there are many manufacturing factors influencing BEF's luminance, in this study we attempt to find out the suitable manufacturing control parameters by employing the hybrid approach of both prediction and optimization techniques. The neural netwo...

Research paper thumbnail of Personal asset allocation using particle swarm optimization approach

Research paper thumbnail of Monthly car sales prediction using Internet Word-of-Mouth (eWOM)

2017 IEEE International Conference on INnovations in Intelligent SysTems and Applications (INISTA), 2017

Internet's mature has changed the behavior of consumers. Most of consumers before purchase wi... more Internet's mature has changed the behavior of consumers. Most of consumers before purchase will queries opinion on the Internet. Vehicle is high priced and durable merchandise, so consumer would be more prudent to view Internet opinion before they buy. Past research has pointed out eWOM (electronic Word-of-Mouth) and customer satisfaction will influence purchasing decision, 70% of consumers believe eWOM. In this study, we utilizes economic indexes, Internet “word of mouth”, and Google Trends variables created by key word searches to forecast the monthly sales volumes of a given model of car. The experiment results demonstrate that the proposed GA/KNN model has the highest predictive power in terms of Mean Absolute Percentage Error (MAPE).