AI requires facility in Linear Algebra,ANN, Calculus ,Probability and RP(Bayesian rules), DSP, Image processing ,NLP and programming. What would be the bare essentials(foundation) from these fields which can be selected so as to ensure sustainable self learning . All are welcome to give their suggestions and experience from industry and academia.
Hi Narasim,
UG generally learns vector dynamics, matrix operations and calculus as part of their curriculum in India hence in case if you really want to include Machine Learning and AI introduction, then showcasing the business usability of all these concepts through Machine Learning would be a great choice.
Following are the outline I can think of to get hands dirty. You may want to choose R or Python to have the students work on live small problems.
- BASIC STATISTICS - CENTRAL TENDENCY - DISPORTION - SKEWNESS - KURTOSIS - COVARIANCE - CORRELATION - CORRELATION TEST
- HYPOTHESIS TESTING - T-TEST - Z-TEST - CHI-SQUARE-TEST - ANDERSON DERLING TEST FOR NORMALITY - BRUSH PAGAN TEST - WHITE TEST - DARWIN-WANTSON STATISTICS - F-TEST - ANOVA - MANOVA - ANCOVA
- REGRESSION (VALUE PREDICTION & CLASSIFICATION) - MULTIPLE REGRESSION - FORWARD REGRESSION - BACKWARD REGRESSION - QUANTILE REGRESSION - POISON REGRESSION - MULTIPLE ADAPTIVE REGRESSION SPLINES - DISTRIBUTION ASSERTIVE REGRESSION - PARTIAL LEAST SQARE REGRESSION - LOGISTIC REGRESION - LOGIT MOEL - PROBIT MODEL - TOBIT MODEL - MULTINOMIAL LOGISTIC - RIDGE REGRESSION - LASSO REGRESSION - ELASTIC NET REGRESSION - NEGATIVE BINOMIAL REGRESSION - SUPPORT VECTOR REGRESSION - BAYESIAN REGRESSION
- MODEL DIAGNOSTICS - VALUE ESTIMATION DIAGNOSTICS - CLASSIFICATION DIAGNOSTICS - CLUSTERING DIAGNOSTICS - ASSOCIATION DIAGNOSTICS - CONCEPT DRIFT - COST FUNCTIONS / LOSS FUNCTIONS - OTHER EVALUATIONS - NOVELTY TESTS - OUTLIER BEHAVIOUR TEST
- ADVANCE STATISTICAL ANALYSIS - PRINCIPAL COMPONENT ANALYSIS- ALL VARIANT - KERNEL PCA - SINGULAR VALUE DECOMPOSITION- WITH USE CASE - SIMULTENIOUS EQUATION - EXPLORATORY AND CONFIRMATORY FACTOR ANALYSIS
- CLUSTERING - K-MEANS - ALL VARIANT - HIERARCHICAL CLUSTERING - GMM - FUZZY LOGICS - BIRCH - DENSITY BASED - SOM - OTHERS
- ADVANCE PREDICTIVE MODELLING - NAIVE BAYS- ALL VARIANT - KNN- ALL VARIANT - RANDOM FOREST - DISTRIBUTED RANDOM FOREST - DECISION TREE - ALL VARIANT - SUPPORT VECTOR MACHINE- ALL VARIANT - ADABOOST - CATBOOST - GADIANT BOOSTING MACHINE - - EXTREEM GRADINT BOOSTING - ARTIFICIAL NEURAL NETWORK -
- TIME SERIES MODELLING - SIMPLE MOVING AVERAGE - ASSUMPTIONS OF TIME SERIES AND TESTING - TIME SERIES DEMOMPOSITION - AR MODEL - ARIMA MODEL - ARIMAX - HOLT-WINTER MODEL AND EXPONENTIAL SMOOTING
Dear Kumarjit
Thank you.That was indeed a very useful and valuable contribution.
" students work on live small problems. " . I was looking forward to this motivation.. It would be nice ,if you can suggest
interesting problems that excite students imagination or point to
some simple intuitive and readable sources for this.
Cheers.
Hi Narasim,
Kaggle.com is a great source of data and business problems to tackle. If they participate in competition and win some awards which may be of good amount money as well, which generally drives the students anywhere in the world for some extra pocket money.
If you are really interested then we can setup some time to go through some of the use case to select them which may be apt for the students of PU based on what curriculum you are looking for.
I do corporate training as well apart from being an data science professional and I had designed the same for the students (colleagues) both junior and senior grade.
However as the students are at UG level they need different set of problems.Hence I need to go over the same again.
Hi Kumarjit .
Is it possible first to post a simple problem from your
stock just to get a feel . Look forward to a fruitful interaction.
Rg gives a warning on Kaggle.com . Is it safe?
Thanks
Cheers
Its hard to do meaningful applied AI without good computer skills (i.e. beyond Matlab script programming). So I would recommend making sure that students are familiar with at least one programming language such as Python, C/C++, Java, etc. Remember, most undergraduates will end up in Industry (not Academia) so it is import that they know how to implement and integrate AI systems.
- Badges
- Science topic
- Similar topics
- Engineering
- Electrical Engineering
- Signal Processing
- Image Processing