Used Car Price Prediction Using Artificial Neural Networks and Machine Learning Techniques

Abstract

Regardless of the substantial rise in car use,
newly constructed motor vehicles fail to
attract potential clients due to a variety of
reasons such as prohibitive costs, shortages
of vehicles, budgetary constraints, along
with other variables. As such, the market
for second-hand automobiles is expanding
quickly everywhere, except for the nation of
India, where it remains very young and
strongly controlled by unorganized
vendors. This raises the possibility of fraud
when buying a vehicle that is used. Asking
choosing the consumer and the
merchandiser in order to determine the
price of a previously owned automobile, a
highly precise estimate is required. This
framework creates two models: a
Computerized Random Walden and an
Oversaw learning-based Artificial Neural
Network framework every one of which can
take advantage of the specified transport
data collection.
A wide range of unique features is taken
into account for accurate projections. The
results obtained beat examples that depend
on simple linear models as well as are
consistent with theoretical predictions.
Artificial neural networks (ANNs). The
collection of data for automobiles is used
for evaluating these methods. The results of
the experiments show that the random
forests approach, which achieves a median
absolute error of 1.0970472 and an R2
mistake that is 0.772584, generates the least
amount of failures compared to every other
approach.
Every day, the world and everyone’s
expectations are expanding. Out of all
the anticipation, one of them will buy the
motor vehicle. As no one is able, the cost of
another one at all times, everybody will
eventually buy a used one. However, a
newcomer is unfamiliar with the asking
price of the wanted automobile on the used
car marketplace. Whenever the purchase
price of new cars increased due to increased
technological costs, the economic case for
older cars became more powerful