Loading presentation...

Present Remotely

Send the link below via email or IM

Copy

Present to your audience

Start remote presentation

  • Invited audience members will follow you as you navigate and present
  • People invited to a presentation do not need a Prezi account
  • This link expires 10 minutes after you close the presentation
  • A maximum of 30 users can follow your presentation
  • Learn more about this feature in our knowledge base article

Do you really want to delete this prezi?

Neither you, nor the coeditors you shared it with will be able to recover it again.

DeleteCancel

Make your likes visible on Facebook?

Connect your Facebook account to Prezi and let your likes appear on your timeline.
You can change this under Settings & Account at any time.

No, thanks

Data Mining

No description
by

Amit Shinde

on 8 June 2013

Comments (0)

Please log in to add your comment.

Report abuse

Transcript of Data Mining

Foundation Design Overview Algorithm Results Math Model The factors considered by the algorithm are: Current weather parameters, favourable conditions for the growth of the pests and the historical pest incidences of the pest in that region.

Current weather parameters are compared with the favorable conditions for the growth of the pest. If there is a match, then the same incidences are searched across the historical data. Introduction GROUP ID - 33

Title: Data Mining in EGovernance for Pest Control

Internal Guide : Prof. Pravin Patil

Deshmukh Mayuraj - B80054230
Pawar Ajinkya - B80054309 The purpose was to design a data mining tool which predicts pest infestations in agriculture. The use of the decision module will enable this tool to understand the input statistics and provides the required output. Data Mining
Decision Support Systems
Content Analysis and Indexing
Statistical Databases
Data Warehouse and Repository Purpose: Technical Keywords: Motivation Dependency of Maharashtra on Agriculture
During 2008-09, in Maharashtra, there was severe outbreak of Spodoptera on Soybean crop. (63.24% of sown area was infested and destroyed - 14 lakh hectares). Financial loss incurred was over Rs.1450 Crores.

Increase in Sucides.

In general crop losses in Maharashtra due to pest problems are estimated to be 20-30% every year.

Crop Pest Surveillance and Advisory Project (CROPSAP). We wanted to create a module for the Agriculture Department Of Maharashtra, to identify pest infestation incidences and provide predictions for the same on a real time basis with a scientific approach.

We want to create this module to:
1. Reduce losses incurred to the farmers.
2. Increase in Productivity.
3. Save the Ecosystem from the heavy use of Pesticides. Village Level Scouts One for every 8 villages Agriculture. Supervisor One for every 10 Scouts Data Entry Operator One for every Supervisor
Online Data Entry twice a week State Agriculture Department Village Board Media Farmers FEEDBACK Stakeholder Advisory Boards NCIPM/ CRIDA/ CICR/ IIPR Our Data Mining Module CROPSAP ARCHITECTURE eAgromet Suppose S is a system such that
S= {I, O, F, Fc, Sc}

Where,
I: - Set of inputs,
O: - Set of outputs,
F: - Set of functions,
Fc: - Set of failure cases,
Sc: - Set of success cases. I: - Set of Inputs

I= {Infestation magnitude, temperature, rainfall, humidity, date}

Input to a system will be the magnitude of infestation of each pest and the corresponding climatic factors considering time as a factor. O: - Set of Outputs

O= {Predicted infestation on timeline, effectiveness of pesticide, pest free zones}

Output will be the Probablity distribution of the system based on the historical data of infestation levels and climate. F: - Set of Functions of eAgromet-
Let function F be F = F1, F2….Fn
F = {F1, F2,…, Fn}
Where,
F 1 = Function to get the infestation level
F 2 = Function to set the climatic factors which affect the infestation
F 3 = Function to map the data onto the previous data
F 4 = Function to perform the data mining to give probability of pest occurence.
etc...
Fn..Fm= Module based functions Sc: - Set of Success cases -

Sc = { Sc1, Sc2, …., Scn }
where,
Sc 1 = Data mining prediction is valid
Sc 2 = Pesticide ineffectiveness is true
Sc 3 = Pest free area successfully mapped

Fc: - Set of Failure cases -

Fc = { Fc1, Fc2, …., Fcn }
where,
Fc 1 = Data mining prediction turns out to be invalid
Fc 2 = Potential pesticide predicted as ineffective
Fc 3 = Pest prone area predicted as pest free Limitations/Constraints

If the supplied data is incorrect or minimal amount of information, output will be affected and forecast will not be credible.

While data mining helps to discover patterns and relationships in data, it cannot promise perfect results, cannot explain why an outcome occurs and cannot correct problems in the data. How is the data collected? Timeline graph plotted using the historical data
for Cotton Timeline graph plotted using the historical data
for Soybean The model that is shown in the diagram shows pest infestation levels for each region plotted as red, yellow, purple, and blue lines. The line for each region has two parts:

Historical information appears to the left of the vertical line and represents the data that the algorithm uses to create the model.

Predicted information appears to the right of the vertical line and represents the forecast that the model makes. The vicinity of the matching historical data instances will give us the rising and falling trends of pest occurrence.

The vicinity records, which give us the trends of the occurrence, also ultimately calculates the probability of occurrence in the future.

Falling trends are discarded, and Rising trends are used to calculate the probability. Specifying favorable conditions Set Weather parameters Sample Output Get Pest Occurrence incidences (historical) Data mining techniques provide mechanisms for in-depth analysis of this data.

We realise that, predicting the incidences of pest infestations and effective control of pests will help to improve the productivity and ecosystem; and save money. However the effectiveness of the result will depend on the input data used. Conclusion Pest attacks have resulted in reduction in yield, in soybean and cotton, this has led to many suicides, which is our major concern. Thank you. Dept. of Agriculture's
Data Warehouse Management Tool Our Data Mining Module For the Dept. of Agriculture's
Data Mining Purposes Our Data Mining Module with extended functionality Thus we have added our data mining module in the eAgromet tool, to provide a consolidated, value-added service for the Department of Agriculture. The output of our data mining module is not value-based but it is probability-based.

This is done so, for 2 reasons:

1. It is impossible to predict nature. A probabilistic answer is the best solution to a possibly varying result.

2. We can analyze the rising and falling trends objectively. With the probability based values, Farmers can be made aware of the possibility of the pest occurence by way of mobile communication/ publicity on a real time basis, which is crucial.

This way the farmers and government can prepare themselves for the control-measured activity, which will reduce the losses and improve food security.
Full transcript