Abstract

 The performance of prediction models is often based on “abstract metrics” that estimate the
model’s ability to limit residual errors between the observed and predicted values. However,
meaningful evaluation and selection of prediction models for end-user domains requires holistic
and application-sensitive performance measures. .The parameters of load forecasting were first
linearly varied in which the desired could not be obtained. Now the real parameters are varied
non-linearly which give the way to intelligent load forecasting. The most successful intelligent
techniques in load, we propose a suite of performance measures to rationally compare models
along the dimensions of scale independence, reliability, volatility and cost. We include both
application independent and dependent measures, the latter parameterized to allow
customization by domain experts to fit their scenario. Load is generic term for something in the
circuit that will draw power which can vary widely. The system load is a random non-stationary
process composed of thousands of individual components. The system load behavior is
influenced by a number of factors, which can be classified as: economic factors, time, day,
season, weather and random effects. Load Forecasting can be thought of as the set of
processes, activities, and toolsets used to create predictions to support operational decision
making of various loads. Our results underscore the value of the proposed measures to offer a
deeper insight into models’ behavior and their impact on real applications, which benefit both
data mining researchers and practitioners.

Existing System

 In exist, to applying data mining and machine learning techniques to domains with societal
impact, and, in the process, understanding the gaps
 In Accuracy consumption prediction for individual customers is less studied, both due to the lack
of input data and the limited need for such predictions.
 As a result, prediction models at the consumer level, with more intra-day and seasonal
variability, and for innovative applications, which engage customers in sustainability, have not
been well understood nor a set of performance measures for their evaluation identified.

Proposed System

 The performance of prediction models are often based on “abstract metrics”, detached from
their meaningful evaluation for the end-user domain.
 We suggest that these measures are insufficient to evaluate prediction models for emerging
domain applications.
 The impact of under- and over-predictions can be asymmetric on the domain, and measures
like The root-mean-square deviation or root-mean-square error is a frequently used measure
of the differences between values (sample and Education values) predicted by a model or an
estimator and the values actually observed.
  To measure errors overlooks the frequency with which a model outperforms a baseline model
or predicts within an error tolerance. Reliable prediction is key for certain domain applications.
 They try to predict the future movement of a stock based on past data. Trend analysis is based
on the idea that what has happened in the past gives traders an idea of what will happen in the
future.
 Our results underscore the value of the proposed to measures to offer a deeper insight into
models behavior and their impact on real applications, which benefit both data mining
researchers and practitioners.

Requirements specification

HARDWARE REQUIREMENTS

 Processor : i3 and above

 Hard drive : 500GB and above
 RAM : 4GB and above

SOFTWARE REQUIREMENTS

 Operating System : Linux/Ubuntu

 Front-end : Java ( MapReduce, Hadoop)
 Back-end : Hive
 Java Tool : Eclipse