Вы находитесь на странице: 1из 27

APPLICATION OF MACHINE LEARNING IN

GEOTECHNICAL PROBLEM

Submitted By
ABHINESH PRAKSH SINGH (15114004)
ANKIT PANWAR (15114011)
FORWARDED BY GUIDED BY
Dr. U. K. DEWANGAN Dr. S. K. CHOUKSEY
Head of the Department Assistant Professor
Civil Engineering Dept. Civil Engineering Dept.
NIT Raipur (C.G.) NIT Raipur (C.G.)
CONTENT
 Introduction
 Application of machine learning
 Regression techniques
 Linear Regression
 Program using linear regression
 Advantages and Disadvantages
 Conclusion
Machine Learning
 Machine Learning is the field of study that gives
computers the capability to learn without being
explicitly programmed.
 It gives the computer ‘The ability to learn’ that which
makes it more similar to humans.
 The basic premise of machine learning is to build
algorithms that can receive input data and use
statistical analysis to predict an output.
Why And When Should We Use Machine
Learning?
 With the rise in big data, machine learning has
become a key technique for solving problems in
many areas like finance, healthcare, research
projects ,engineering works etc.
 It is very useful for a complex task or problem
involving a large amount of data and lots of
variables, but no existing formula or equation.
 Machine learning is a good option if we need to
handle situations where the nature of the data
keeps changing, and the program needs to adapt
as in automated trading and predicting trends.
How Machine Learning Works
 Machine learning uses two types of
techniques: supervised learning, which trains a
model on known input and output data so that it
can predict future outputs.
 And unsupervised learning, which finds
hidden patterns or intrinsic structures in input
data.
Regression techniques
 Regression techniques are used when we are
working with a data range or if the nature of it’s
response is a real number.
 Common regression algorithms include linear
model, nonlinear model, stepwise regression.
Linear Regression
 Linear Regression is a machine learning
algorithm based on supervised learning. It
performs a regression task.
 Regression models are target prediction values
based on independent variables.
 Different regression models differ based on –
the kind of relationship between dependent and
independent variables, they are considering and
the number of independent variables being used.
How does it work ?

 Mathematically, regression uses a linear


function to approximate (predict) the dependent
variable given as:
Y = βo + β1X + ∈
where, Y - Dependent variable
X - Independent variable
βo - Intercept
β1 - Slope
∈ - Error
βo and β1 are known as coefficients. This is the
equation of simple linear regression.
 Error is an inevitable part of the prediction-
making process.
 We can't completely eliminate the (∈) error
term, but we can still try to reduce it to the
lowest.
 To do this, regression uses a technique known
as Ordinary Least Square (OLS).
 OLS is easy to analyze and computationally
faster, it can be quickly applied to data sets
having 1000s of features.
 Interpretation of OLS is much easier than other
regression techniques.
 In OLS, the error estimates can be divided into
three parts:

 Residual Sum of Squares (RSS) - ∑[Actual(y)


- Predicted(y)]²

 Explained Sum of Squares (ESS) -


∑[Predicted(y) - Mean(ymean)]²

 Total Sum of Squares (TSS) - ∑[Actual(y) -


Mean(ymean)]²
Program-1
 We have created a random data-set to train our
model.
 The plot for the data set generated using the
code is shown below.
 The model parameters are given below.

 R^2 score is near to unity so it will provide


satisfactory results.

 RMSE is the square root of the average of the


sum of the squares of residuals.
 The plot of the best fit line is shown below.
Program-2
 Now we have used this technique in slope
stability analysis of infinite slope with the given
set of data which includes one dependent
parameter ( Factor of Safety FOS) and five
independent parameters,
Angle of slope(β), Angle of internal
resistance(Ф), Cohesion of soil(c), Bulk unit
weight of soil(ᵞ), depth of failure surface below
ground(z).
 The plot for the data set generated using the code
for FOS and cohesion(c) shown below
 The plot for the data set generated using the code
for FOS and β shown below
 The plot for the data set generated using the code
for FOS and Ф shown below
 The plot for the data set generated using the
code for FOS and ᵞ is shown below
 The plot for the data set generated using the
code for FOS and z is shown below
Advantages of Machine Learning
Language
 Easily identifies trends and patterns
 No human intervention needed (automation)
 Continuous Improvement
 Handling multi-dimensional and multi-variety
data
 Wide Applications
Disadvantages of Machine
Learning
 Data Acquisition
 Time and Resources
 Interpretation of Results
 High error-susceptibility
Future scope of machine
learning
 Machine Learning (ML) or Artificial
Intelligence (AI), are almost used as synonyms
for each other and are perhaps the most talked
about topics in recent times.
 The global machine learning as a service market
is rising expeditiously mainly due to the Internet
revolution.
 The process of connecting the world virtually
has generated vast amount of data which is
boosting the adoption of machine learning
solutions.
CONCLUSION
 We have used Root mean squared error(RMSE)
and Coefficient of Determination(R² score) to
evaluate our model.
 For more accurate pridiction R² value should be
near to unity.
 In case of slope angle and depth of slice the R²
value is greater than 0.8 it shows that these two
parameters have greater weightage on FOS than
other two values.
References
 https://en.wikipedia.org/wiki/Linear_regression
 http://machinelearningmastery.com/simple-linear-regression-
tutorial-for-machine-learning
 Nabil, K. Neural networks in civil engineering II: Systems and
application. // Computing in Civil Engineering. 8, 2(1994), pp. 149-
162.
 ANG, A.H.-S.; TANG, W.H. (1975): “Probability Concepts in
EngineeringPlanning and Design, Vol. I: Basic Principles”. John
Wiley and Sons, NewYork.
 BAECHER, G. B. (1982): “Statistical methods in site
characterization”.Updating subsurface samplings of soils and rocks
and their in-situ testing, Engineering Foundation, Santa Barbara,
California, pp. 463-492.

Вам также может понравиться