Required knowledge:

Knowledge of statistics, finances, and business administration

Objectives, learning outcomes and obtained knowledge:

Predictive Ananlysis of time series. Mapping problems in predictive analytics, solutions in practice.   Support by  standardized tools. Show and understand the surplus of visualization, and turn it back to the data preparation and modeling phases

Synopsis :

Modul_1.: Visual analytics

Introduction to Predictive Analytics and Visualization, visual analytics

Analytical reasoning techniques,

Data representations and transformations

Visual representations and interaction techniques.

Generalized multidimensional scaling

Perceptual mapping

Business Decision Mapping (BDM)

Practice in Laboratory 1: Visualization

Modul_2: Forecasting

a.)   Approaching a forecasting problem

Components of a time series; Judging the quality of data; Understanding data; Looking at residuals; How to start making a forecast; Forecasting models.

Defining parameters, Analysis of data sources; Choosing alternative projection techniques  Preliminary selection criteria

b.)   Forecasting with exponential smoothing models

Smoothing with moving averages; Single exponential smoothing; Compare exponential smoothing with moving averages; Exponential smoothing for trending data

Practice in laboratory 2.  Exponential smoothing; Software programs and visualization.

c.)   Trend and seasonality modeling and analysis;

ANOVA model; Contribution of trend/seasonal effects;  Analysis of residuals.

Practice in laboratory 3: Trend and seasonality; Software programs and visualization

d.)   Preparing the data for modeling;

Achieving linearity; Achieving normality; Dealing with outliers

Practice in laboratory 4: Outliers; Software programs and visualization.

e.)   Regression modeling and analysis

Building regression models: The regression curve; A simple linear model;  The method of least-squares; Normal regression assumptions; Comparing estimation techniques;  Interpreting regression output: The R-squared statistic; The t-statistic; The F-Statistic; The D-W Statistic; Assessing forecast precision, Looking at regression residuals.

Practice in laboratory 5: Regression example; Software programs and visualization.

f.)    Insuring against unusual values

The need for robustness in correlation and regression analysis Seasonal adjustment; Ratio-to-moving-average-method. Seasonal adjustment with resistant smoothers

Practice in laboratory  6: with seasonality analysis; Software programs and visualization

Modul_3: Foresight

a)     Differences of foresight and forecasting

Non measurable trend analysis: qualitative description, success factors Topic definition, starting position Ongoing projects, expected development Visualization of trends through drawing, pictures (like cicles, hype)

b)     Visioning a usage area

Topic definition, summary of the situation Driver analysis estimation of effects, uncertainty Scenario making, alternative scenarios, illustrations Visualization and illustration of the visions

c)     Technology radar for foresight Flow of news, scanning news, practice for selections Professional blogging, technology radar  Virtual community to build up Games for knowledge integration

d)     Strategy making based on backward scenarios

Choosing objectives, freedom of choices, views Influencing drivers, costs and risks  Strategy forming through backward scenario analysis

Practice in laboratory 7: Foresight presentations in of the students on a preliminary given topic

Summary: Usability of predictive analysis, foresight and  visualization.

Method of instruction :

Lectures and 6 practices in laboratory

Assessment:

a. In the class period there is 1  in-class test (ZH) from the topics of modul1 and modul2

                                   1 written and presented homework  from the topic modul3

b. In the examination period: written  examination and it could be extended orally,

c. Preliminary examination opportunity exists

d. Condition for the signature is the pass mark of ZH test minimum 4 points from the maximum 10 points. Another condition for the signature is at least successfull attendances the laboratory exercises. One practice in laboratory can be missing.

Recaps :

There is one possibility to repeat the test in the teaching period. In the rectification period(repeat period) there is another (final) possibility to rewrite the in-class test (ZH).

Only two of practices in laboratory can be repeated in an appointed time with the instructor.

The homework presentation can be repeated int he recap period in a given data, with paying the recap fee.

Consultations:

Preliminary appointing with instructors or after the lectures

References, textbooks and resources:

1. Box G., Jenkins G. M., Time Series Analysis – Forecasting and Control, CA: Holden-Day, 1976.
2. Sallehuddin, R., Shamsuddin, S. M. H., Hashim, S. Z. M., Abraham, A.: Forecasting time series data using hybrid grey relational artificial neural network and auto regressive integrated moving average model. 2007.
3. Wong P C, Thomas J.: Visual analytics. 2004.
4. Mark Last, Abraham Kandel, Horst Bunke: Data Mining In Time-Series Databases. World Scientific Press. 2004.
5. Liam Fahey (Editor), Robert M. Randall (Editor): Learning from the Future: Competitive Foresight Scenarios (1997)
   

Syllabus prepared by:

Name	Possitions	Department
KÓSA Zsuzsanna Ph.D	Associste professor	Department of Telecommunications and media-informatics