Abstract

Owning to the booming of social media, making comments or expressing opinions about merchandises online becomes easier than

before. Data from user reviews might be one of the essential inputs for forecasting sales of vehicles. Besides, some other effects, such as

stock market values, have inuences on purchasing power of vehicle. In this project, both multivariate models with user review data and

stock market values and time series models are employed to predict monthly total vehicle sales. Three types of data, namely sentiment

scores of user reviews, stock market values, and hybrid data, are employed in this project to forecast monthly total vehicle sales. The

hybrid data contain both sentiment scores of user reviews and stock market values. In addition, seasonal factors of monthly total vehicle

sales are employed to deseasonalizing both monthly total vehicle sales and three types of input data. Thus, both user review data and

stock values are essential to forecast monthly total vehicle sales; and deseasonalizing procedures can improve forecasting accuracy in

predicting monthly total vehicle sales. Online reviews provide consumers with rich information that may reduce their uncertainty

regarding purchases. As such, these reviews have a significant influence on product sales. In this paper, a novel method that combines the

sentiment analysis while using historical sales data and online review data is developed for product sales forecasting. A sentiment analysis

method, the Naive Bayes algorithm, is used to extract the sentiment index from the content of each online review and improve the

forecasting accuracy.We collected real-world automotive industry data and related online reviews. The computational results indicate that

the combination of sentiment analysis has higher forecasting accuracy than the some other sales forecasting models.
 Objective
Consumers tend to spend a considerable amount of time searching for information about a potential vehicle.
Sales forecasting in the automotive industry is particularly important since in the current system are either
built-to-delivery or built-to-forecast. However, the latter one often leads to a bullwhip effect due to uncertainty
in demand and inaccurate forecasting. Even if product are built-todelivery, accurate forecasting can still help
managers to plan and allocate their resources better. User reviews act as word of mouth and allow companies
to collect large-scale and up to date data that represents honest consumer opinions. A popular method used by
companies to analyze these data is sentiment analysis, which analyzes people’s opinions, sentiments,
evaluations, attitudes, and emotions from natural language. Understanding these sentiments is important for
product and service evaluation and these sentiments also may have an impact on people’s future purchases.
 Literature Survey

 Title: Sentiment Analysis and Google Trends Data for Predicting Car Sales

 Author: Fons Wijnhoven, Olivia Plant

Consumers tend to spend a considerable amount of time searching for information about a potential vehicle. The study showed that a majority of
customers spend more than 10 hours to identify the best vehicle for their requirements. In China, this number even reached 70%, while in western
countries like Germany and the USA, a percentage of 40-50% stated that they spend at least this amount of time. Consequently, search volumes —as
registered by search engine firms like Google.com— may represent a large part of people’s searches for decisional information for their car purchases.
Therefore, previous research has studied the predictive power of Google Trends data for car sales, although with different success. The best car sales
prediction models combine Google Trends data, Forum sentiment and Forum mention volumes as predictors. In their study, Google Trends has a similar
predictive power as Forum sentiments and volume together. Our aimed at contribution here is a further identification of the meaning of these predictors.

 Title: Product sales forecasting using online reviews and historical sales data

 Author: Zhi-Ping Fan, Yu-Jie Che, Zhen-Yu Chen

Firms use product sales forecasting as a foundation to estimate sales revenue and make decisions regarding production, operation and marketing
strategies. Through product sales forecasting, firms can create a plan for marketing, sales management, production, procurement, logistics and so on to
improve their economic benefits and reduce losses caused by weaknesses in the production plan. According to the extant research, two primary factors
influence consumers' purchasing decisions. One is the influence of other consumers who have bought the product and recommended it through verbal
communication. The other is the influence of advertisements and the mass media, among other factors. A number of researchers have studied product
sales forecasting and developed effective forecasting models that take relevant factors into account
Literature Survey
 Title: A Study of Sentiment Analysis and Sales Prediction: Tourism Domain

 Author: Odilia Gonsalves

In this study, We aim to predict sales for a tourism sector based on the mined reviews. To validate the forecasting model, real world data from a tourism firm is

used. The algorithm for sales prediction is validated and compared against other time series models. The performance of Holts Winter method is compared with the

performance of other time series techniques. Results obtained from sales prediction show that Holts Winter Method performs better than the other time series

forecasting techniques. The sales prediction is calculated as a combination of sentiment analysis score and Holts winter method. The derived result states that the

forecasted value is varied based on the sentiment analysis score for past months. The conclusion is that for a tourism sector having seasonal trend, Holts Winter

Method in combination with the predictive power of reviews performs better than the other time series forecasting techniques. In future, the same data can be used

to predict customer behavior and utilize it for improving the companys business can be implemented. Also different algorithms like ARSA can be applied on the

current dataset and their accuracy in forecasting can be calculated.

 Title: Learning to Learn Sales Prediction with Social Media Sentiment

 Author: Zhaojiang Lin, Andrea Madotto, Genta Indra Winata, Zihan Liu, Yan Xu

Social media sentiment has shown to be a useful resource for product sales forecast. However, research on modeling the correlation between sentiment index and

sales is often limited by the scarceness of quarterly sales data. In this paper, we propose to learn how to learn sentiment-sales correlation from different source

products and transfer to sales prediction of another, target product. We evaluated our approach on sales data of seven different smartphones and showed that the

knowledge transfer from six source products significantly reduced the sales prediction error for the target product, in a 7-fold cross-validation experiment.
EXISTING SYSTEM

 Before the Internet was developed, transmission of information is mainly through ways, such as leaets, billboards,
television and word of mouth. Recently, the Internet has been growing; and restrictions of space and distance have been
reduced. Information of products can be updated at any time by social media. Businesses provide more real-time and
convenient services and closely interact with consumers. Due to the rapid ow of messages, the electronic word of mouth
and brand image is deeply influenced by social media; and purchase willingness is affected as well.

 In existing aspect identification techniques uses aspect extractor for identify aspects in new reviews. For this task Hidden
Markov Model and Conditional Random Fields approaches have been used. Supervised learning techniques depend on
train data set. This approach is reasonably effective, but preparation of training example is time consuming. In
unsupervised approaches automatically extract product aspects from consumer reviews, without using training example. In
this approach focus on association rule mining based on the Aprior Algorithm to mine frequent item sets as frequent
product aspect.
DISADVANTAGES OF EXISTING SYSTEM:

 cannot obtain precise results

 syntactic patterns are prone to errors. online reviews usually have informal writing styles, including
grammatical errors, typographical errors, and punctuation errors. this makes the existing parsing
tools, which are usually trained on formal texts such as news reports, prone to generating errors.

 time consuming because of manual process

 less efficient

 opinion of sentiment word would be content sensitive

Online Word of mouth, such as online reviews and microblogs, have become popular with the
development of Internet technologies. In Proposed system we are using navie bayes classification
for sentiment analysis of user reviews for forecasting sales of vehicles. the data set of independent
variables including sentiment scores of user reviews and two stock market values of vehicles are
processed. Sentiment analysis refers to extracting sentiments and opinions from user reviews text.
Common problems with identifying and classifying sentiments as positive, negative or neutral
arise when the author expresses multiple opinions. In this project, a method that combines the
sentiment analysis is proposed to forecast product sales using product review data. This method
incorporates the Naive Bayes (NB) algorithm to compute the sentiment index of online reviews.
The system will analyze the comments of various users and will rank the product finally. The
database of the system has various keywords denoted as positive and negative words, that helps the
system to identify the comments and rank the products accordingly. The system receives the
comments from the users, based on the comment, system will specify whether the comment is bad,
better or good. This system helps the users to select the correct product for an individual.
 
 ADVANTAGES
 • precise results can be obtained
 • errors free.
 • Less Time consuming because of computerised process
 • More efficient
Tools
 
Hardware Requirements:
 
System: Pentium IV 2.4 GHz.
Hard Disk : 500 GB.
Monitor : 14’ Colour Monitor.
Mouse : Optical Mouse.
Ram : 4 Gb.
Keyboard: 101 Keyboard.
 
 
Software Requirements:
 
Operating system : Windows 7.
Coding Language : PHP
Data Base : MYSQL.
Modules
Admin Module
Admin Login
Admin logs into the system by specifying unique username and password.
Add products
After login admin can add the product details like category, name, description, product image, price and quantity.
These details gets stored in product table.
Admin View Review Opinion words
In this module admin can view the no.of reviews added to the product. Admin can also view the product
classification whether product falls into positive Opinion words or negative Opinion words.
 Admin View Review Opinion Targets
 In this module admin can view the no.of reviews, no.of Positive Opinion, no.of Negative Opinion added to the
product. Opinion Targets are classified based on rule
 Positive Opinion Targets=(No. of Positive Opinion ) / (Total no.of reviews)
 Negative Opinion Targets=(No. of Negative Opinion ) / (Total no.of reviews)
 If Positive Opinion > Negative Opinion, the product is classified as good product
If Negative Opinion > Positive Opinion, the product is classified as bad product.
Logout
After view Opinion Target reviews of the product, admin can now logout from the session.
User Module

User Registration

The System requests for user name and password. User must enter the correct values and submit. Now the system checks for sub code format and existence,
then it will store the user’s information. The System informs the user as Information was saved.

User Login

User need to login into the system If user not yet having username and password to login, they need to register by giving their basic details and get username
and password.

View Product and review Product

After logging into the system, user can now view the shopping products online. Now user can add reviews to the product. There are two types of Opinion
words dataset classified, one is positive Opinion words and another one is negative Opinion words.

User reviews are compared with the Opinion words dataset and classified using navie bayes classification algorithm, whether it is Positive Opinion or
negative Opinion. User can review the individual product only once, if he review same product twice a ‘Reiview already added to this product’ error alert
will be shown to user.

Logout

After adding reviews to the product, user can now logout from the session.