Hi,   Can u suggest any books or online courses in order to prepare for the SAS Base exam ? (besides the preparation guide) I'm specifically looking for actual exam questions samples and exam simulations, in which the answers are explained. Furthermore, what is your experience regarding the workload? I'm working full-time, so I'm not sure how much weeks/months/weekends I need to spend. However, I have already some SAS experience since I attended university lectures. What is your experience?    Thanks, KS ... View more

Hi, I have a general question on text mining regarding a business case. Note sure if I understood the concept of text mining 100% but let me explain my issue: Imagine you need to analyze and categorize insurrance contract agreements as pdf-files. For example the contract contains a statement like: "cancelation of the contract is possbile after death" or "cancelation of the contract is possible whenever you like". In that case I would categorize both statements into the "cancelation" categories plus a furhter subcategory needs to be created for "cancelation -> death" and "cancelation -> free".   Imagine now, another sampled contract would contain a sentence like "the termination of the contract is only possbile when the agent passed away". (sorry couldn Thing about something better :P).  In that case it would fit to subcategory death but only content wise. So far so good, I guess manually creating your dictionary would be possible.   My question is now: Is there any automatized/ ML-learning approach to categorize this classification dictionaries without manually defining and categorizing sentences? Maybe it is possible to "train" a dictionary ML algorithm?   And in case, can the SAS Text Mining Software solve such Tasks? Thanks, KS ... View more

Im thinking of applying machine Learning algorithms on a data set of a Portfolio of credit contracts. The dataset is huge. I got around 22 Million contracts with around 100 Million data rows from a 2000 to 2015 (Panel Data) and up to 30 individual characteristic variables. With this data set I want to estimate the so called "prepayment risk". Generally speaking, I have a regression problem where I want to estimate the parameters in order to forecast the probability of prepaying a loan within a group of clients In the literature such models are basically estimated with a logistic Regression because the dependend variable is usually discretized. With some extended things were also modelled in a survival Analysis modell. The goal of my research should be, how or if neural networks can improve the estimation compared to a logistic regression. Hence I want to get a step further and estimate my parameters with a deep learning ANN. Unfortunately I don't have any experience with machine learning, but I'm a graduate student of econometrics. Now I want to read some opinions from some "experts". Do you think this is a good idea for a master thesis? I'm afraid that in the neural Network Approach I might get stucked at some point. Could you give me some advise about some good papers and books? Maybe some online lectures which are worth to watch? Can these things be done in SAS? ... View more

Hi,   is there any SAS procedure/method in which I can estimate a multinomial logit model (or something similar) on panel data? So far I just know that it is possible to run a fixed-effect logistic regression with use of the STRATA specification but as far as I know this only works for binary choice models. I'm open for other techniques like GMM Estimation or even Machine Learning algorithms. Any suggestions?  Thanks, KS. ... View more

Hi, I have a dataset with prepayment data,specifically CPR I got high N, but relatively short T. Additionally my data set is highly unbalanced. I want to test the panel for autocorrelation & unit root. Durbin-Watson test does not apply since it only works for balanced panel data.   My model looks as follows:   proc panel data=.. ... model cpr = spread loan_age loan_age2 rgdp CPI /unitroot(..); Unfortunately nothing seems to work. Additionally, i tried all of the following tests manually: BREITUNG, COMBINATION (or FISHER), HADRI, HT, IPS, and LLC But either my panel was too unbalanced, nor it told me that i dont have enough observations( eventough i have around 50,000 N), nor I got an error message that I am using too many lags(?). The Breitung tests even output me "floating point error"?. Does someone have an idea or an alternative how i could fix this problem? Maybe there is another way of checking for autocorrelation & unit root of my panel? Thanks, KS ... View more