Sea Shells Study for Information Gain and Analysis

trusciencetrutechnology@blogspot.com

Sea Shells Study for Information Gain and Analysis with the object of assessing the marine information content of its species in the Indian Ocean

by

Tallamraju Satya Sandeep

8^th Std Student, Ramakrishna High School, Maharanipeta P.O, Visakhapatnam. 

Res:16-3-4, s/oTallamraju Surya Bhagawanulu,

Official Colony, Maharanaipeta.P.O. Visakhapatnam-530002

Phone: 08912508152

Introduction

It is a well established study the Information Technology, for induction of decision trees using the gain. Below such an approach except for the tree presentations that has been postponed for future investigations by us, I present the manner in which the sea shells provide excellent and very significant information gain by way of its attributes.

OBJECT OF STUDY

I have collected Bay of Bengal Sea shells near the RK Beach shore in Visakhapatnam with the object of assessing the Marine Information content of its species in the Indian Ocean. The idea that the Bay of Bengal has an influx from the Indian Ocean of Marine life of minutest nature is thought because the Bay of Bengal Sea offers itself as a safe haven for the marine small creatures due to it being basically a trough bay of life and hide for survival. The migratory nature of these species and development of off springs is a subject of outstanding information technology possessed by the marine creatures especially the small organisms. It is well known the tortoises travel several kilometers if not thousands of miles to get their food stuff jelly fish and the marine small organisms. The information content in a very cursory way has been modeled and analyzed by me with the intention like studies would become very popular and would guide the information technologists of marine environment studies. This no doubt forms a new and novel subject by itself and our attempt is just the first step.

THE METHOD

The method adopted by me is very simple and straightforward for practical applications. First A Table is created with the thumbnail pictures of the sea shells that are randomly collected to provide a best possible sample. The day and season of the collection is of course obviously important and I have chosen the month of November when the Bay of Bengal Sea would be highly turbulent and both earthquake and tropical storms prone. The washed out miniature organisms shells would thus give a representation of both the very far distant off shore and the depth of the sea marine organisms. I have carefully collected about 33 shells of different types which is indeed a random collection. It may be noted that these are left behind by the shell trading community fishermen people at the off shore beach who thrive on a colored and modeled shell structures of various designs to attract the tourist population to the Beach. Hundreds of local people make a simple living by it. So the collection has the uniqueness of the missed minute sea shells that have not attracted their attention for trade benefits. Thus true marine organisms were collected. The broken and distorted sea shells is blessing in disguise since they carry with the Information Gain of their attributes.

ATTRIBUTES

There are several attributes that serve the purpose of information gain analysis such as their color, strips, banded and vertical line or curve features. Also the dots and the mixed color attributes may be analyzed. I have first used the Excel to list these and make the information gain calculations. Simple classes of m1 and m2 are identified first of all with the label attributes of shine or no-shine. This served the purpose of information needed to classify a tuple in say D. Information gain is defined as the difference between the original information requirement (i.e.based on just the proportion classes) and the new information obtained after partitioning the split sub set say A. That is to say we calculate the difference to obtain the Gain of the split data. This helps to normalize the Gain calculations in general and in particular the split information as well. The method thus becomes very clear by presentation of these calculations made by me and the results presented in the following section. The partition technique is very helpful to design analyze the information hidden in the Sea shells of the Marine environment. The obvious physics in it is a far reaching finding.

33 sea shells collected by Sandeep from RK Beach Visakhapatnam

INFORMATION GAIN CALCULATIONS

Color                     Strips                      Shine                 color y/n
brown 1 light     no                             yes  1                   y1
light yellow 1    2 bands                                   no 1                    y2
white  1                vertical 6 bands                   no   2                  n1
light brown 2     spiral bands         yes   2                  y3
orange  1             circular 5 curves no   3                    y4
white   2               vertical 23                no  4                 n2
light yellow 2    bands two               yes   3                y5
green  1                circular bands  6    yes  4                y6
white   3vertical lines 7+bands 7   no  5                  n3
red   1                    curves 12             no   6         y7
gray  1                   5 bands                                yes 5         y8
white   4               1 band                  yes   6       n4
light yellow  3      no                        no   7        y9
light brown  3      no                         no   8         y10
red         2              spiral bands       no   9     y11
light green 2      yellow dots        no  10    y12
light yellow 4    2 bands                                no  11    y13
gray   2                  2 bands                                yes  7     y14
gray  3                   5 bands                                yes  8     y15
light yellow 5    no                           no  12    y16
Gray 4                   no                           no  13    y17
light brown  4    14 bands              no  14    y18
white   5               7 curves               no  15    n5
RED 3 black yellow          dents    no   16   y19
white   6               3 bands                                no   17   n6
red         4              no                           no  18    y20
White 7                                6 bands                                no  19    n7
Red 5                3 broad bands    yes  9   Y21
red    6                   spiral bands       yes  10  y22
gray    5                 no                           yes  11  y23
red dots   7          12 curves             yes   12 y24
red     8                  25 curves             yes   13 y25
gray    6                 no                           no  20   y26
infoDshellshine=-13/33*log2(13/33)-20/33*log2(20/33)			class m1=shine
infoDshellshine =	0.9673		class m2=no shine
infoDshellcolor=-26/33*log2(26/33)-7/33*log2(7/33)
infoDshellcolor =	0.7455	bits
Gain in information from such a partitioning is 0.9673-0.7455----------------------->
infoDshellbands= 4/33*(-3/4*log2(3/4)-1/4*log2(1/4))+5/33*(-2/5*log2(2/5)-3/5*log2(3/5))+7/33*(-6/7*log2(6/7))+1/33*(-1/1*log2(1/1))+2/33*(-1/2*log2(1/2)-1/2*log2(1/2))+8/33*(-7/8*log2(7/8)-1/8*log2(1/8))+6/33*(-3/6*log2(3/6)-3/6*log2(3/6))
infoDshellbands                                         =	0.6601
light yellow 5
white 7
ornange 1
green 2
red 8
gray 6
total=4+5+7+1+2+8+6=33 Gain= InfoDshell-infoshellbands color Partition =0.9673-.6601=0.3072 bits
Sandeep on Theerdhapu Raallu of RK Beach Visakhapatnam

ACKNOWLEDGMENT

I am thankful to my father Shri Tallamraju Surya Bhagawanulu, Retired UNI Senior Correspondent, for encouraging me to do this project work.