Menu
Home
Log in / Register
 
Home arrow Computer Science arrow Applied Big Data Analytics in Operations Management
Source

PROS AND CONS FRAMEWORK

Opinion forms the basis for people, organizations and social communities to take accurate and effective decisions. People mostly ask their friends, peers and knowledgeable persons about their opinion on an entity while taking the decision, since, they believe the experiences, observations, concepts and beliefs of other individuals will help them in boosting the decisiveness of that entity. With the recent growth of the social media on the web in the form of reviews, blogs, Twitter, forum discussions and social networks, the inception of expressing the opinions and views on these portals has become a huge trend. Due to this large volume of opinionated data is being loaded for analysis. Thus, opinion mining plays a pivotal role to focus on the opinions which express or imply positive or negative sentiments. It involves both Natural language Processing(NLP) and machine learning techniques. Opinion Mining has various applications. It can be applied to different consumer products and services. It can be used for twitter analysis and measuring sales performance of an organization.

It can be categorized into three levels-Document level, Sentence level and Aspect level. The Document level classifies the review into either positive or negative as it assumes that whole review is about the single entity. The assumption considered in this level is that it is being assumed that the whole document consists of an opinion about the same or single entity. For example-a review about a particular product whether it is a camera, phone etc.. In such types, opinion about the particular product is calculated in terms of either positive or negative as a whole .However, it will not be applying for a blog post because in such posts the opinion holder compares different products with each other.

The Sentence level describes the sentence into subjective or objective by using subjectivity analysis and then the opinion orientation of each sentence is found. This level is closely related to Subjectivity Analysis by(Tsytsarau et al., 2011). In this, the first task is subjectivity classification where a sentence is categorized into objective or subjective sentence. Objective sentences are those sentences which contain facts or no sentiments explicitly. On the other hand, subjective sentences are those which contain opinion or sentiments . However, it must be noted that the objective sentences may also contain opinions implicitly. For example- consider the sentence“The battery backup of the camera is very great”. This sentence has an opinion defined explicitly so it is a subjective sentence. Now consider “The screen broke in two days”. It appears that the sentence is objective but actually it provides an important opinion about the screen of the phone. Thus, subjectivity classification poses as a challenge task. The second task is the sentiment classification which aims at finding the opinion orientation of each sentence by classifying them into positive, negative or neutral sentiment.

The aspect level is quite popular among the researchers. It determines the most important features of an entity by extracting the aspects which describe the entity. The different approaches of opinion mining are Sentiment Classification, Subjectivity Analysis, Lexicon based, Statistical based, Dictionary based and Semantic based. For example-“This room is very large”. This sentence has an aspect “room” of the entity “hotel”. The orientation of the opinion is positive . An another example can be cited from the movie domain as” I didn’t like the storyline yet the movie was scintillating”. In this, although the orientation of “storyline” is” negative” but the “movie” has an orientation “positive”. Feature level allows to classify different features into different polarities-mainly positive and negative.

The different approaches of opinion mining are Sentiment Classification, Subjectivity Analysis, Lexicon based, Statistical based, Dictionary based and Semantic based. Sentiment classification is the technique of classifying the polarity with supervised learning.

The assumption considered in this level is that it is being assumed that the whole document consists of an opinion about the same or single entity. For example-a review about a particular product whether it is a camera, phone etc.. In such types, opinion about the particular product is calculated in terms of either positive or negative as a whole .However, it will not be applying for a blog post because in such posts the opinion holder compares different products with each other.

Sentiment Analysis or opinion mining is the field of analyzing social media text using machine learning or NLP techniques. With the growth of Web 2.0, the data on social media platforms is increasing in huge amounts. This data is in the form of text and is highly unstructured. There comes the importance of big data. This huge volume of data, when analyzed, can be used by supply chain managers in forecasting the sales of the products and creating marketing strategies. It can be done by analyzing the social media postings of different customers about a product. The different pros and cons of the application of opinion mining in operation management is described in the following sections.

The Figure 1 describes the different advantages of applying opinion mining in operation management.

Advantages of Opinion Mining in Operation Management

The advantages of the application of opinion mining in operation management are listed below:

Figure 1. Pros of opinion mining in operation management

Forecast Actual Sales:Research and validations show that the sales of an organization can be predicted by the analyzing the social media behavior. (Abbasi et al. 2012).It has been investigated that there is some sort of relationship between the products reviewed online, the release of the products and the actual sales of products . This relationship can provide an aid in predicting the sales of the products.(X. Yu, Liu, Huang, & An, 2012) .This prediction is important in the overall growth of the organization. The sales of the product is related to the overall revenue of the organization.

  • • Identify Different Volumes of Customers: The sentiment analysis, based on the underlying social media reviews about different entities can be used to provide an easy method for an organization to search different customers in each segment. Customers at different levels have unique things to say about the product. This pipeline is described with the fact that the customers who are interested and have potential may ask about the attribute of the product .The owners of the product can make comments about the product either positive or negative. There are some ‘unaware customers’ who do not know anything about the product. There is another level of consumers who are aware of product but not interested. These type of customers ask general questions about the product. The Figure 2 shows the above concept in a pictorial manner.
  • • Appreciation of consumer is evaluated while new initiatives are formed:

A consumer response to supply chain sustainability initiatives(SCSI) can help a firm to gain customer response using opinion mining(Wood 2014b). Firms can get the idea from the opinion mining whether the changes made in products lead to more revenue or not. If consumers are incurious about the new initiatives then organizations must not introduce new SCSI.

• Profitable for Distant Firms: Opinion Mining helps the distant organizations to take better decisions in understanding the consumers . It opens the avenues for these firms located far away from their customers to cater their need for better and quality products. Opinion mining combined with operation management can deal with different volumes of consumers in different streams. The comments of the consumers are elucidated to indicate the decisions that are needed to be taken by an organization.

Figure 2. Pipeline of consumers (Warren 2008)

Disadvantages of Opinion Mining in Operation Management

There are certain limitations of applying opinion mining in sentiment analysis which are illustrated below in Figure 3.

Miscategorised Post Dealing with Sarcasm

A major limitation of opinion mining is the text to be dealt with sarcasm and irony. Not much attention is given to the posts that contains sarcasm as these types of posts generate a wrong value of sentiment. The incorrect value of the sentiment can lead to inaccurate information about the customers’ sentiments of the products in an organization. This misleading and miscategorised post can hamper the growth by decreasing the sales and revenue of a firm.

 
Source
Found a mistake? Please highlight the word and press Shift + Enter  
< Prev   CONTENTS   Next >
 
Subjects
Accounting
Business & Finance
Communication
Computer Science
Economics
Education
Engineering
Environment
Geography
Health
History
Language & Literature
Law
Management
Marketing
Mathematics
Political science
Philosophy
Psychology
Religion
Sociology
Travel