The Study on Relationship Verification of e-wom and Hotel Operation Performance by Data Mining Technique

The power of e-word-of-mouth are impact Hotel industry dramatically in the recent years. The Ambassador Hotel Group is a five-star rated business hotel type in Taiwan. Quantitative and qualitative data are retrieved from internet respectively based on data mining technique. Descriptive and regression analysis approach are used for quantitative data, meanwhile, text analysis and words meaning assortment are used for qualitative data. Result demonstrated that higher internet exposure rate will impact hotel’s operating performance positively. ‘Services’ is most important factor toward reviewer’s perspective regarding hotel impression. Otherwise, ‘Services’ is most comment word appeared in the reviews. Some suggestions are mentioned base on research findings.


Hotel background
Ambassador Hotel Group was established on December 1, 1962, to response the call of the government to promote the tourism industry with a project of the premier international tourist hotel construction-Ambassador Hotel Taipei in capital of sixty million New Taiwan Dollars and in 1982 was listed on the Taiwan Stock Exchange [1].
As the leading brand of Taiwan's five-star tourist hotel, the Ambassador has sustained to explore the market of Taiwan to support the tourism industry. The Ambassador emphasized the integrity, teamwork, innovation, and the spirit of contribution to society as their principle of operating to business [2]. In addition, the Ambassador also get involved in the business of hairdressing salon, laundry service, swimming pool operating, car park, sauna, GYM room and store renting etc. [2].

Data mining
"Data mining" introduced by Frawley, Paitetsky-Shapiro and Matheus of the term as early as 1991, which is defined as "…is the nontrivial extraction of implicit, previously unknown, and potentially useful information from data." [3,4] raised a similar definition to the data mining, "Data mining is the application of specific algorithms for extracting patterns from data". In recent years have seen the growing of the technology of data mining research [4] deemed that data mining is an investigation of the valuable data from a database and widely used in the field of data processing [5] deemed that the application of the system of information technology should makes user receive the information according to their requirement instead of others useless information.

Mouth marketing
Word of Mouth is a straightforward but informal way of communication in the human society, which could influence consumer behavior in either short or long period [6,7]. Since the internet becomes more popularity, the effect of word of mouth is no longer confined in the way of communication by face to face. Internet has given an opportunity for consumer to share their user experience which these information and reviews could be receive by one another consumer as a first-time user and further influence their purchasing decision, is referred to as e-WOM [8]. Usually, Word of Mouth Marketing emphasized on transaction between market and consumer, in fact, there is concerning about the value of goods or services bring to consumers well. E-WOM marketing believe that consumer will not necessarily to consume from the reason of good reputation in WOM, but a good reputation in WOM will make consumer accelerate in purchasing decision making and more influence their original opinion. To the consumer, company, goods or services with a good reputation in WOM will more easier to make consumers purchasing and coming back for it [9].

Data collection
To verify the authenticity of analysis results for triangulation which including quantitative data and qualitative data. All data used in this study were retrieved from multi-resources and based on data mining technique. First, the data of financial statements (URL: www. ambassadorhotel.com.tw/financial/financial.htm) and the internet search results by term of "Ambassador Hotel" are collected through Google search respectively. Hence, the relationship between operating performance [10] and internet exposure rate are examined. The collected data are divided in quarter from 2008 to 2013 (Table 1).
Second, the e-WOM indicators are retrieved and weighted from TripAdvisor (URL: www.tripadvisor.com.tw), which is the famous website of consumer-generated reviews on hotel and catering in Taiwan [11]. The data are ranged from 2006 to 2013 (Table 2). All data are collected manually by researchers from internet on December 12th to 16th, 2013.

Analysis
A multi-analysis approach are used in this research [12]. The descriptive analysis and simple regression are used for quantitative data analysis. Text analysis and words meaning assortment are employed as research approach for qualitative data.

Results
The characteristic of sample As shows in Table 3, the type of Ambassador Hotel [13] is classified as business hotel. Ambassador Hotel Hsinchu is the newest compared with other two Ambassador Hotels in Taiwan. All of the belonged hotels are 5 stars rated by Taiwan Tourist Hotel Association. The largest capacity of hotel rooms is Ambassador Hotel Kaohsiung, but Ambassador Hotel Hsinchuis the smallest one. The features of each hotel are different: Ambassador Hotel Hsinchuis positioned as Luxury Business Hotel [14]. The average operation income of Ambassador Hotels were NTD. 696,844,000 in quarter (Table 3). Table 4, there is a significant relationship between operating performance (dependent variable) and internet exposure rate (independent variable) with an Adjusted R² of 28.3%. The result supported that the internet exposure rate can predict the Ambassador Hotels operating performance positively [15]. The detailed functions as following.

The review of e-word-of-mouth recommendations
The purpose of reviewers to dwell in the Ambassador Hotel Hsinchuis mainly for business trip, only few reviewers for leisure tour. All text recommendations are reviewed and 8 keywords to be noted. The most comment word is "Services". The words "Environment and facilities", "Foods and beverages", and "Location" are also to be promoted by reviewers. Meanwhile, "Price" and "Traffic" are slightly needed to be promoted compared with other words.

Conclusions and Suggestions Conclusions
Ambassador Hotel Hsinchu is set as "Luxury Business Hotel" and located at Hsinchu, the part of northern Taiwan. The researchers found a positive relationship between operating performance and internet exposure rate. It demonstrated that higher internet exposure rate will impact hotel's operating performance positively.
From the reviews of Hotel Rating from Trip Advisor indicating that the "Services" is the most important factor which impact reviewer's satisfaction toward Ambassador Hotel Hsinchu.
On the other hand, empirical findings of e-WOM analysis indicated that reviewers are satisfied Hotel's sanitary and cleanliness, Booking and check-in services, and Business services. An inconsistence results appeared compare with Hotel Rating. Customers are satisfied their services provided by Ambassador Hotel Hsinchu. This point shall be further investigated to discover the actual reason.

Suggestions
Some empirical suggestions are provided based on research results for hospitality industry as supporting information for decision making. Obviously, necessary specified strategies shall be provided to promote internet exposure rate. This action can enhance operating performance of Ambassador Hotel.
Further, "Services" and "Price" shall be improved immediately to motivate reviewers' impress regarding to the Ambassador Hotel Hsinchu. In addition, the weakness of hotel location shall be improved by enhance transportation services to promote the reviewers' satisfaction. Otherwise, customer's interview shall be conducted to understand the reason of inconsistence between Hotel Rating and e-WOM.
In summary, more positive e-WOM recommendations can facilitate Hotel's operating performance. This finding complied with recent researches [11,15,17]. Therefore, Hotel executives shall assign importance to e-WOM and give some supports to reviewers on related websites.