UNIDEP

Professional Development University

Master's degree :

Financial administration

Subject:

Analytics and data mining

Job:

Video report: Predictive Techniques for modeling

of data.

Teacher:

Rafael Antonio Rivera Cruz

Student:

Brisa del Carmen Cárdenas Hernández.

Date of delivery:
February 19, 2023
 INDEX

INTRODUCTION...................................................................................................................3
VIDEO REPORT....................................................................................................................4
CONCLUSION.....................................................................................................................12
BIBLIOGRAPHY.................................................................................................................13
 INTRODUCTION
Predictive techniques are a set of data analysis tools and models used to predict
the future behavior of data. These techniques are used to predict the trend of data,
predict demand for products or services, and predict user behavior. These
techniques are especially useful for making decisions in managing a business.
Predictive techniques are based on the collection of data in a structured way and its
subsequent analysis. Predictive techniques can also be applied for data mining and
network analysis. These techniques are used to identify trends, patterns and
relationships between different data sets. These techniques are used to better
understand user behavior and to identify business opportunities.
Predictive techniques can also be applied for fraud detection. These techniques are
based on identifying anomalous patterns in the data that may indicate fraudulent
behavior. These tools can also be used to detect possible cyber attacks.
These tools are based on data collection and subsequent analysis to train
predictive models. These models allow predicting the future behavior of the data.

Within this work, a report of six related videos was made on the topic of predictive
techniques for data modeling.
 VIDEO REPORT
VIDEO 1

Predictive modeling: design and implementation process, testing, validation of the

results and selection of the most beneficial prediction model, which is supported by
different analysis methods and learning logarithms.
Validation of the prediction model : ensures that you have the most appropriate
model to predict the probability of a result.
The modeling methods available in predictive analytics software platforms are:
Statistics, data mining, machine learning and artificial intelligence.
The prediction model is selected based on criteria of testing models, validation and
evaluation of results based on the detection theory and assumption of probability of
an outcome given a given data sample.
The models can use different logical mathematical classification techniques to try to
determine the probability that one set of artistic data or another variable takes a
certain value, which is known as prediction.
The predictive model available in the predictive analytics software platform can be:
• Sales forecasts based on time series.
• Disease detection.
• Wholesale price
• Fraud detection system.
All these predictive models allow us to obtain new information from the input data
and thus select the predictive model that best suits the business. Each model has
its strengths and weaknesses and is designed for particular types of problems.
A predictive model is reusable and is created by training an algorithm using
historical data and saving the model for later uses by sharing general business
rules that can be applied to similar data in order to analyze the results using the
trained algorithm and new ones. data.
Standardized predictive modeling processes that are critical in the predictive
analytics software platform are:
1 .- Creation of the model: includes the design and implementation of prediction
models that will be trained from one or more algorithms taking the training data
from the sample.
2 .- Model testing: using learning algorithms and sample training data in some
scenarios tests are carried out based on historical data to observe and confirm the
best prediction of the model.
3 .- Model validation: validates the results of the model through understanding of
business data and visualization interface.
4 .- Model evaluation: allows the evaluation and selection of the prediction model
that best fits the historical results. The modeling process involves the iterative
execution of one or more algorithms on the sample data. This process is known as
the model training name.

CATEGORIES OF PREDICTION MODELS

Prediction models on the other hand classified as:
Predictive models : These analyze historical data to make future predictions.
Descriptive models: analyze historical data and establish relationships between
the data with the purpose of classifying data sets into groups.
Decision models : also known as prescriptive models. Establishes the
relationships between the various elements of a decision to predict its results and
select a decision.
https://youtu.be/TLmCVItX5xw
 VIDEO
2
DATA MINING IN PREDICTIVE ANALYTICS
Data mining: is the process of discovering behavioral patterns of data in large
volumes with the help of computer and applications of artificial intelligence,
machine learning, statistics and data management system. The purpose of data
mining also known as KDD or Discovery data base is to convert large volumes of
unorganized data into structured information for further analysis.
Data processing: is a multidisciplinary analysis process that must be carried out by
statistical mathematicians and computer programmers.
Data mining encompasses various processes, including: collection, extraction,
storage, cleaning and standardization, transformation, analysis, statistics, artificial
intelligence, machine learning and intelligent business. Before the data is explored,
that is, subjected to the statistical analysis process, it needs to be cleaned and
transformed to eliminate errors and inconsistencies.
Statistics provides tools for analysis and establishes patterns and trends in data,
and machine learning manages the different learning methodologies.
Statistical analysis techniques are: mathematical procedures that seek to establish
patterns and trends or individual or group reflections. The most important statistical
analysis methods or techniques include generalization, characterization,
classification, clustering, association, evolution, sequential patterns, decision trees,
data visualization and rule-guided mining.
The data mining process does not stop and continues after a solution learned
during the process was obtained.

https://youtu.be/rv8Y-jpRFao

DATA ANALYSIS IN PREDICTIVE ANALYTICS

Data analysis in predictive analytics : inspection process, cleaning,
transformation and data modeling with the purpose of finding useful information,
reaching conclusions and making appropriate decisions. In statistics, data analysis
is divided into: descriptive analysis, exploratory analysis and confirmatory
analysis. , also known as deductive or inferential analysis.
The data must be clean, so outliers, closed values and unwanted values must be
 VIDEO
3
eliminated. These cleaning processes will vary depending on the type of data, for
example, for quantitative data, the detection of outliers eliminates anomalous
values. Meanwhile, spell checkers can be used to reduce misspelled words in the
case of text-type data.
Business intelligence is a process of data analysis based on aggregates,
disaggregates, averages, segmentations, etc., all of these focused business
information.
Predictive analytics is the application of statistical models for predictive forecasting
to determine events that will happen in the future.
Text analytics: it is the application of statistics. linguistics and structural models to
extract and classify information from texts and be able to interpret it for prediction
purposes.
Analysis methods
Data analysis is composed of various phases:
PHASE 1: INITIAL ANALYSIS
Initial analysis, data cleaning, quality of values, data quality.
PHASE 2: MAIN ANALYSIS
Exploratory and confirmatory analysis, stability of results, statistical analysis
https://youtu.be/FQig0WIO8_8

Data analytics framework To start working with data analytics for a company, you
must establish a framework that consists of 3 types of analytics.
The first of them is descriptive analytics, the second is predictive analytics
and the third is prescriptive analytics.
In descriptive analytics, the first thing we arrive at and have to know is what is
the structure of the data we have, what information we have, what information we
validate and can work with and what information is not suitable for working with
them, the ideal is to start structure which part of the service processes I have to
start digitizing in order to start collecting data from said processes.
Predictive analytics is where, once we have counted the data, we begin to work
 VIDEO
4
on it and begin to develop, in this case, predictions about the possible results that
can be obtained, then we begin to have the equivalent of a batting percentage and
a reaction percentage. What does such a hitter have when faced with another
situation, then we are already beginning to have, well in this case, now we do have
models that help you predict and anticipate the results that are going to occur.
Prescriptive analytics , when I start working on the models I can start exchanging
variables to start developing their position on how I can improve and have the best
performance on my decision making.

https://youtu.be/4F8G87Ho7SI
 VIDEO
5
DATA MODELING IN BIG DATA
It helps transform a large volume of data into useful information for business
development through data analysis, which is the process of inspecting, cleaning,
transforming and modeling data.
The modeling phase aims to build an analytical model from the data table that has
already been generated.
An analytical model is essentially a mathematical formula to which you enter data
and will return a result.
The stages of this modeling phase are the following:
• Know the typology of modeling technique.
• Define the evaluation metric to be used.
• Train or build the models.
• Evaluate the final model.
Initially, the modeling techniques are discussed, which are grouped into typologies,
the most important being two, supervised learning and unsupervised learning.
Unsupervised learning problems are those where you don't have a lot of data and
you ask the algorithm to find or identify patterns without giving you any clues.
In supervised learning, those clues are delivered based on existing data from which
the algorithm will learn; however, in supervised learning, those clues are delivered
based on existing data based on existing data with which the algorithm will learn.
Examples of supervised algorithms would be: linear regression, logistic regression,
decision trees and neural networks. Once the learning typology is clear, the
evaluation metric that allows measuring the capacity of the model must be defined.

https://youtu.be/pZeYbv35g2g
PREDICTIVE ANALYTICS
Predictive analytics : it is the branch of advanced analytics used to make
predictions about future events that companies will face, for which they use various
techniques or methods such as data
mining, text mining, statistics, modeling,
machine learning. and artificial
intelligence.
Predictive analytics is the process of
information technology management,
analysis methods and business process
modeling with the purpose of ensuring
predictions about future events that
companies will face.
Structured business data such as sales
reports, country information, employee, supplier, product, region, etc. come from
our moderated databases and can be used for analysis directly, while unstructured
data such as those come from Word documents, PDF content from social
 VIDEO
6
networks, emails, pages, games, blog, sensors, call center note or other types of
open texts, they need to be extracted from the texts to infer the user's feelings and
preferences and can be used in the model building process.
Forward-looking proactive predictive analytics anticipates results based on data
and not assumptions, predictive analytics such as customer relationship
management through marketing campaigns, customer service, preventive
maintenance, in the areas technical service, etc. In the health sector through
disease detection systems and determination of medical treatment. In the financial
sector for the detection of fraudulent transactions, identification theft and false
insurance claims. In risk management it can be used to predict the best portfolio to
maximize return.

Predictive analytics is a complex process of anticipating events that will happen in

the future and can be carried out through complementary or parallel processes that
help the final objective, which is to foresee risk situations and future events that
companies will face.
Predictive analytics processes are:
Project definition : this defines the results deliverable to the business objective,
data that will be used in the project.
Data collection and mining : connect and prepare data through data mining that
comes from various sources that will be used in the prediction model. Data mining
is the computer-based process of data patterns, trends and behavior from
enterprise big data, with the help of machine learning technologies,
Statistics, the objective of data mining is to extract information from big data and
convert it into structured information for analysis through predictive analysis
models.
The data mining process is an interdisciplinary process that requires the
participation of statistical mathematicians and computer programmers. Data
analysis : it is the process of review, cleaning, transformation and statistical
analysis of data base with the objective of raising hypotheses and demonstrating
them and thus obtaining useful information and reaching conclusions that allow
decision making.
Statistical analysis : it is of greater depth than the previous analysis and allowed
the assumptions and hypotheses to be validated and tested using standard
statistical models, the statistical analysis begins with the determination of the
population and the process to be analyzed, the population is a set of observations
of the process in different periods of time, the statistical analysis can be descriptive,
in this case the population is described through aggregates that are shown in
graphic form.
Predictive modeling : it is the process of design, implementation, testing and
validation of a prediction model, the models are chosen through the testing,
validation and evaluation processes to detect the degree of reliability in the results.
A model is reusable and is created based on a training algorithm using historical
data and storing the model for later reuse in similar processes. The general
processes of predictive modeling are: design, implementation of the model, testing
of the model, validation of the model and model selection.
Implementation: allows the implementation of the results of the selected prediction
model through decision automation, the validation and implementation of predictive
models are time-consuming activities that can take months and depend on the
characteristics of the business scenario.

https://youtu.be/Nb3VCmFLipA
 CONCLUSION

Predictive techniques for data modeling are a useful tool to help industry

professionals make better-informed decisions based on data analysis. These

techniques can be used to predict the future behavior of data, to improve the quality

of products and services, to improve the profitability of the company and to make

strategic business decisions. Experts in the field of data modeling can help industry

professionals identify and exploit opportunities for improvement, as well as identify

trends and patterns in data to predict future results. These techniques are a key

tool for making informed decisions.

 BIBLIOGRAPHY
AddKw. (December 2, 2018). Retrieved from https://youtu.be/FQig0WIO8_8

AddKw. (December 8, 2018). Retrieved from https://youtu.be/rv8Y-jpRFao

AddKw. (January 10, 2019). Retrieved from https://youtu.be/TLmCVItX5xw

AddKw. (JANUARY 19, 2019). Retrieved from https://youtu.be/Nb3VCmFLipA

FUENTEVILLA, N. (MARCH 25, 2021). COOL IDEAS. Obtained from

https://youtu.be/4F8G87Ho7SI

VALUE WORLD. (OCTOBER 4, 2021). Obtained from

https://youtu.be/pZeYbv35g2g