Servlet technology is used to create web application (resides at server side and

generates dynamic web page).

Servlet technology is robust and scalable because of java language. Before Servlet, CGI
(Common Gateway Interface) scripting language was popular as a server-side
programming language. But there was many disadvantages of this technology. We have
discussed these disadvantages below.
There are many interfaces and classes in the servlet API such as Servlet, GenericServlet,
HttpServlet, ServletRequest, ServletResponse etc.

What is a Servlet?
Servlet can be described in many ways, depending on the context.

Servlet is a technology i.e. used to create web application.

Servlet is an API that provides many interfaces and classes including

documentations.

Servlet is an interface that must be implemented for creating any servlet.

Servlet is a class that extend the capabilities of the servers and respond to the
incoming request. It can respond to any type of requests.

Servlet is a web component that is deployed on the server to create dynamic web
page.

Text Representation

Features BoW

TW OA

K P R F1

Dp -

69.9% 21.6%

37% 29% 33%

BoW binary

72.9% 28.8%

69% 36% 48%

BoW tf-idf

73.8% 30.0%

75% 38% 50%

BoW+Dp binary

73.8% 30.0%

73% 38% 50%

First Level Classification

Second Level Classification

BoW+CF+Dp tf-idf

BoW+Dp tf-idf

80.0% 48.1%

75.7% 35.0%

76% 59% 66%

74% 37% 49%

BoW+CF binary

78.7% 46.5%

74% 58% 65%

BoW+CF tf- idf

79.4% 46.4%

71% 54% 61%

RESULTS FOR THE TWO STAGES OF THE PROPOSED HIERARCHICAL CLASSIFIER

Text Representation First Level Classification Second Level Classification
Features BoW TW OA K P R F1
Dp - 69.9% 21.6% 37% 29% 33%
BoW binary 72.9% 28.8% 69% 36% 48%
BoW tf-idf 73.8% 30.0% 75% 38% 50%
BoW+Dp binary 73.8% 30.0% 73% 38% 50%
BoW+Dp tf-idf 75.7% 35.0% 74% 37% 49%
BoW+CF binary 78.7% 46.5% 74% 58% 65%
BoW+CF tf-idf 79.4% 46.4% 71% 54% 61%
BoW+CF+Dp binary 79.1% 48.3% 74% 57% 64%
BoW+CF+Dp tf-idf 80.0% 48.1% 76% 59% 66%
TABLE III
AGREEMENT BETWEEN FIVE EXPERTS ON MESSAGE NEUTRALITY
Classification Neutral Non-Neutral
Expert OA K P R F1 P R F1
Expert 1 93% 84% 97% 93% 95% 97% 93% 95%
Expert 2 92% 80% 91% 98% 94% 95% 78% 85%
Expert 3 95% 90% 99% 94% 97% 88% 99% 93%
Expert 4 90% 76% 89% 98% 93% 94% 73% 82%

Expert 5 94% 84% 94% 97% 95% 93% 85% 89%

TABLE IV
AGREEMENT BETWEEN FIVE EXPERTS ON NON-NEUTRAL CLASSES IDENTIFICATION
Violence Vulgar Offensive Hate Sexual
Expert P R F1 P R F1 P R F1 P R F1 P R F1
Expert 1 89% 99% 94% 89% 97% 93% 80% 90% 85% 78% 98% 87%
Expert 2 77% 83% 80% 92% 67% 78% 71% 60% 65% 71% 69% 70%
Expert 3 81% 84% 83% 76% 96% 85% 67% 79% 72% 53% 89% 66%
Expert 4 96% 41% 58% 92% 78% 84% 70% 60% 65% 79% 42% 54%
Expert 5 84% 90% 87% 92% 77% 84% 77% 73% 75% 78% 84% 81%

0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%

TABLE II
RESULTS OF THE PROPOSED MODEL IN TERM OF PRECISION (P), RECALL (R) AND F-MEASURE (F1) VALUES FOR EACH

82%
85%
84%
97%
85%

98%
67%
76%
64%
77%

89%
75%
80%
77%
82%

CLASS

First level Second Level

Metric Neutral Non-Neutral Violence Vulgar Offensive Hate Sex
P 81% 77% 82% 62% 82% 65% 88%
R 93% 50% 46% 49% 67% 39% 91%
F1 87% 61% 59% 55% 74% 49% 89%

First level
Neutral

Non-Neutral

Second Level
Violence Vulgar Offensive Hate
Sex

Metric
P
R
F1

81%
93%
87%

77%
50%
61%

82%
46%
59%

62%
49%
55%

82%
67%
74%

65%
39%
49%

88%
91%
89%