About Krishika

By 

Altamash Haider Khan

In the Indian society agriculture plays a very important role. According to 2011 agricultural Census of India, an estimated 61.5% of the 1300 million Indian population is rural and dependent on agriculture. The number offarming households are 159.6 million. The % of marginal households with less than 1 ha of farm land has come down from 62.9% in 2000–01 to 22.5% in 2010–11.

Climate plays a dominating role in agriculture. Plants require sufficient heat and moisture for their growth. Normally, regions having maximum temperature of less than 10°C are not suitable for plant growth. In the tropical regions, where temperature is high throughout the year, agriculture is successfully done.

Plant life is not possible in dry areas except that with the help of irrigation. The moisture requirements vary from plant to plant and region to region. In the lower latitudes, where temperature is high, plants need more moisture for their growth (75cm to 100cm).

On the other hand, in the higher latitudes wheresummers are cool, winds are not dry, rainfall of 50-62 cm is sufficient for plant growth.

So, we thought how useful it would be if we applies the machine learning  to predicts the weather features so, farmers could know which crop they need to cultivate and what precautions they need to take in case of any mishappening.

Project also predicts the expected production in terms of kilogram per hector unit. This further helps in the cost analysis.

FEASIBILITY STUDY

Feasibility study is conducted once the problem is clearly understood. Feasibility study is a high level capsule version of the entire system analysis and design process. The objective is to determine quickly at a minimum expense how to solve a problem. The purpose of feasibility is not to solve the problem but to determine if the problem is worth solving.

The system has been tested for feasibility in the following points.

1.  Technical Feasibility

2.  Economical Feasibility

3.  Operational Feasibility.

1.  Technical Feasibility

The project entitles "KRISHIKA” is technically feasible because of the below mentioned feature. The project was developed over web technologies.

In this we had to implement machine learning algorithms to predict the crop production for the upcoming years.

2.  Economical Feasibility

This project is economically really very feasible as it is developed over web with minimum requirement of resources. It will direct its audience as farmers who will be benefitted really well economically because it will help them to predict the production for upcoming years.

3.  Operational Feasibility

The project is operationally sound and not too much difficult to understand and implement. This is developed to be used by not to well educated farmers so major focus is given on more number of images rather than words.

ANALYSIS

3.1Problem definition

1.      Data cleaning:

Data cleansing or data cleaning is the process of identifying and removing (or correcting) inaccurate records from a dataset, table, or database and refers to recognizing unfinished, unreliable, inaccurate or non-relevant parts of the data and then restoring, re-modelling, or removing the dirty or crude data. Data cleaning may be performed as batch processing through scripting or interactively with data wrangling tools.

After cleaning, a dataset should be uniform with other related datasets in the operation. The discrepancies identified or eliminated may have been basically caused by user entry mistakes, by corruption in storage or transmission, or by various data dictionary descriptions of similar items in various stores.

2.      Predicting future climate conditions:

In project five features have been predicted which are precipitation, Average Temperature, cloud cover, vapour pressure and reference crop evapotranspiration. This has been done using the exponential running average.

Exponential moving averages (EMAs) reduce the lag by applying more weight to recent data. The weighting applied to the most recent price depends on the number of periods in the moving average. EMAs differ from simple moving averages in that a given day's EMA calculation depends on the EMA calculations for all the days prior to that day. You need far more than 10 days of data to calculate a reasonably accurate 10-day EMA.

There are three steps to calculating an exponential moving average (EMA). First, calculate the simple moving average for the initial EMA value. An exponential moving average (EMA) has to start somewhere, so a simple moving average is used as the previous period's EMA in the first calculation. Second, calculate the weighting multiplier. Third, calculate the exponential moving average for each day between the initial EMA value and today, using the price, the multiplier, and the previous period's EMA value. The formula below is for a 10-day EMA.

3.      Predicting yield of crop per unit hector

When the future data regarding the climate conditions have been predicted, the next ask is to predict the yields of crop per unit hector. This task is done using the multivariate regression in which total 20 features have been used to predict the yield of the crop. This data can give.