You should learn to use R and its libraries since this is free. Be sure to check answers for some simple text book data sets. In commercial software,  I find minitab very user friendly for beginners. JMP is also easy to use. I do not know what is emg data.

I recommend R (https://www.r-project.org/) as well! Install R and Studio (https://www.rstudio.com/) and your life will be much easier. The learning curve is a bit slow at the beginning but once you get the hold of it, it is the best software tool I have used.

Not sure what emg data is (electromyography?) but I am 99.9% sure R will be able to handle it quite neatly. ;)

Yes, EMG stands for electromyography.

basically I want carry out principal component analysis and anova test on recorded data

I know nothing about EMG data, but would also highly recommend R.

I have not found that the R leaning curve is any steeper than that of Stata for instance.

I have done both PCA and anove in R and it is not particularly difficult. You do want to spend a little time however, researching which package to use for PCA.

I recommend R, SAS or GENSTAT; the advantage of using R is that it's free, the others aren't.

Also would recommend R for the reasons already mentioned. Whatever you are looking for, chances are there is a package available for it. It is more than just a statistical package but a programming language. For analysis of EMG data, for example there is a package 'biosignal EMG''.

thank you all. i look forward to know more suggestions.

what about XLTSTAT?

R

You can use MATLAB, R, SCILAB, SPSSS, PYTHON etc..

i am also collecting data from EMG and use SPSS to analyze my data, ex: independent-test, anova, regression, etc...

But, if your data is big i suggest using Matlab, i am also use it for my big data. Because i record EMG data for 45 minute, For 1 sc i have 1500 data., and you can imagine the total of my data for 45 minute :D

SPSS is easier for beginner, no need to use programming language...

You'll find that R is essentially the standard and cutting edge in much of the sciences, in particular academia. And every day people are developing new uses and packages for things from new statistical techniques and big database access to GIS and mapping uses. While I'm unfamiliar with your specific needs it's well worth giving it a go, regardless of it being free!  

I agree with Scott about R. Also, there is an online course on coursera.org about using R in statistics offered by Brian Caffo. that might be of help. 

https://www.coursera.org/learn/r-programming 

I mainly use SPSS in my works however. It is limited when you are using the interface but if you know Python you can quickly develop codes and modify them based on your needs. So, it is easier to use compared to R initially. But if you want to do advanced statistical coding/analyses the limitations will bother you unless you code in Python. 

I would also suggest R. It is hard to learn, often frustrating. But eventually it is worth it!!

While learning R you will also have to learn a bit "statistical thinking", you will have to learn more about the statistical methods and the structure of data. As being a programming language you will have to learn programming what will train and structure your way of thinking. This is an additional complication - but again: it is absolutely worth it!

Here are some helpful links to learn R:

https://r-dir.com/

http://tryr.codeschool.com/levels/1/challenges/1

https://www.datacamp.com/courses/free-introduction-to-r

https://stat545-ubc.github.io/topics.html

https://www.safaribooksonline.com/library/view/getting-started-with/9781449314798/ch01.html

Further I strongly recommend using RStudio as user interface to R:

https://www.rstudio.com/products/rstudio/download/

What kind of EMG software that you used?I use Noraxon software for analyzing the raw data then use SPSS for the statistical analysis... SPSS is more user-friendly for me compare to others...

Here you can find tons:

https://cran.r-project.org/other-docs.html

Perhaps starting with "R for beginners" by E. Paradis