Anda di halaman 1dari 28

Know your Future

Machine Learning, AI and Predictive Analytics


Demo 1

Loading Huge Volumes of Data into a Big Data cluster

Automatic Exploratory Data Analytics using FinLibre Stack

Querying Data at Speed


The Overall Architecture
Let’s See a Demo to load several
million records in sub seconds
Automatic Exploratory Data Analysis

Use full in Brining up the dash boards Quickly.

A Starting point for understanding the data.

Helps in More detailed investigations.

No Coding and No SQL, any one can do.


Let’s see a Demo of
Automatic Data Exploration
Demo 2

Time Series Forecasting using ARIMA

Time Series Forecasting using


Prophet
ARIMA
• An ARIMA model is a class of statistical models for analyzing and
forecasting time series data.
• Explaining seasonal patterns in sales
• Predicting the expected number of incoming or churning customers
• Estimating the effect of a newly launched product on number of sold
unit.
• Detecting unusual events and estimating the magnitude of their
effect
Let’s Predict Stock Prices
Prophet
• hourly, daily, or weekly observations with at least a few months (preferably
a year) of history
• strong multiple “human-scale” seasonalities: day of week and time of year
• important holidays that occur at irregular intervals that are known in
advance (National Holidays)
• a reasonable number of missing observations or large outliers
• historical trend changes, for instance due to product launches or logging
changes
• trends that are non-linear growth curves, where a trend hits a natural limit
or saturates
Predicting the Book Sales of
Spiritual Publishing House
Demo 3

Regression and Prediction, Types of


Regression

Classification, Clustering, Identifying


Patterns in Data.
Machine Learning

Machine Learning: The use of algorithms that can learn from and make
predictions on data, as opposed to traditional programing with
instructions.
• Related to Data Mining, Pattern Recognition, statistical learning, etc.
• Supervised Vs Unsupervised Learning
• Supervised Learning: Task of creating a function/relationship from training data which
has one or more explicit output (dependant) variables. Also, indicated as data that is
labelled. This can then be used to map new instances of the inputs.
• Unsupervised Learning: Task of creating patterns from data which have no explicit
measure or signal guiding us. The data is unlabelled
• Semi-supervised Learning
Machine Learning

In Supervised Learning
• Problem type: Regression versus Classification
• Objective: Prediction versus Interpretation
Unsupervised Learning
• Clustering
• Association Rule Mining
How Prediction Works
Regression and Prediction
Regression Analysis is the study of relationships between variables
• Going beyond categorical variables
• Model based relationship (linear and non-linear)
• Useful towards interpretation and prediction

Examples:
• Relationship between speed and fuel efficiency of a Truck; Operating drill speed and soil/stone
density affect on rate of drill
• How do wages of employees get affected with experience, education, promotions, etc.
• How does sales revenue get affected as a function advertising expenses, competitors
advertisements, etc.
• How does the price of a property get affected by square footage, distance from city centre, etc.
• How does the current price of a currency exchange rate depend on its past values
Categorizations, nomenclature and concepts
Linear versus non-linear Simple versus
multiple

Cross-sectional versus time-series

Response (or Dependent) variables

Explanatory (or Independent) variables


Scatter plots and outliers

Correlations
• Regression and Prediction
Regression and Prediction in action

Simple linear regression

Multiple Linear Regression


Regression
Live Demos
• Understanding Example Using Excel
• ATM Withdrawal Prediction
• Small Loan Default Prediction
Classification – Demo 4
• Classification algorithms are used when the desired output is a
discrete label.
• Classification and class probability estimation attempts to predict, for
each individual in a population, to which class does this individual
belongs to.
• Predicts Several things line
• Loan and credit card defaults
• Type of Customer (frequent, Spends more, brand conscious)
• Product Classification
• Logistic Regression, Support Vector Machines, Decision Trees and
Random forest
Let’s See the demo for
• Lets Predict Loan Defaulters
• Credit Card Defaults
Pattern Analysis and Clustering – Demo 5
• Identify Hidden Patterns in Data.
• Understanding customer behaviour
• Market Basket Analysis
Let’s See the demo for
Customer Pattern Analysis
Market Basket Analysis
K-Means Clustering (Jupyter Notebook and Anaconda)
Model Comparison
• Model Comparison Manual Method
• Model Comparison Automatic Method (h2o)
ML Where to Start
• R – Lang (R Studio)
• Swirl - Learn R, in R.
• Jupyter (Python)
• Tensorflow for Deep Learning
• But Before all the above
• High school Mathematics
• Linear Algebra
• Matrices
• Probability theory and Statistics
Thank You

Anda mungkin juga menyukai