Naive Bayes Classifier in Python

Selection_005
Standard

Naive Bayes Classifier is probably the most widely used text classifier, it’s a supervised learning algorithm. It can be used to classify blog posts or news articles into different categories like sports, entertainment and so forth. It can be used to detect spam emails. But most important is that it’s widely implemented in Sentiment analysis. So first of all what is supervised learning? It means that the labeled training dataset is provided along with the input and the respective output. From this training dataset, our algorithm infers the next outcome to a given input.

The basics,

Conditional Probability : It is simply the probability that something will happen, given that something else has happened. It’s a way to handle dependent events. You can check out some examples of conditional probability here.

So from the multiplication rule; (here A and B are dependent events)

P(A ∪ B) = P(A) · P(B|A)

Now from above equation we get the formula for conditional probability;

P(B|A) = P(A ∪ B) / P(A)

Bayes’ theorem : It describes the probability of an event based on the conditions or attributes that might be related to the event,

P(A|B) = [P(B|A) · P(A)] / P(B)

So, our classifier can be written as :

Assume a problem instance to be classified, represented by vector x = (x1, x2, x3, …. , xn) representing some n attributes. Here y is our class variable.

Selection_005

Here we have eliminated the denominator P(x1, x2, x3, …. , xn) because it doesn’t really contribute to our final solution.

Now to make sure our algorithm holds up good against our datasets, we need to take the following conditions into account.

The Zero Frequency problem : Let us consider the case where a given attribute or class never occurs together in the training data, causing the frequency-based probability estimate be zero. This small  condition will wipe out the entire information in other probabilities when multiplied (multiplied by zero…duh…!). The simple solution to it is to apply Laplace estimation by assuming a uniform distribution over all attributes ie. we simply add a pseudocount in all probability estimates such that no probability is ever set to zero.

Floating Point Underflow : The probability values can go out of the floating point range hence to avoid this we need take logarithms on the probabilities. Accordingly we need to apply logarithmic rules to our classifier.

I have implemented Naive Bayes Classifier in Python and you can find it on Github. If have any improvements to add or any suggestions let me know in the comments section below.

image4233

Refer :

Flickr Wallpapers daily for Linux Mint & Ubuntu

best-free-photo-sharing-websites-like-flickr-300x251
Standard

Linux has very few options for ‘Automatic Wallpaper Setter’. Variety is at the top. I loved it, it has lots of image sources, effects, a digital clock and daily quotes. You can also set your wallpaper-switching interval. It is really worth checking out. But what I wanted was a simple script that would download images from Flickr daily and set them as my desktop wallpaper. So here is the Python script to download and set wallpapers on Linux Mint and Ubuntu from Flickr based on an image’s interestingness.

So, to get a list of interesting photos from Flickr we will require Flickr API keys (used strictly for non-commercial use). You can get them from here. The API we are going to use for our script is : flickr.interestingness.getList more info about this API can be found here. You can set its arguments as per your requirements. For this script, I am requesting the response in XML format but you can also get it in JSON. When we are done with tuning the API it would look something like this.

https://api.flickr.com/services/rest/?method=flickr.interestingness.getList&date=2016-04-11&per_page=5&api_key=YOUR_API_KEY

(You need put your API key there…). The above REST request doesn’t actually return an image URL but it returns Image IDs, Server IDs and Secret alphanumeric IDs of interesting images as per your limit (in this case 5) and date. Now our job is to map these image IDs to their respective static URLs so that it will be convenient for us to download them.

https://farm{farm-id}.staticflickr.com/{server-id}/{id}_{secret}_[mstzb].jpg

With the help of the above API, we can construct the source URLs of the images. At the tail of this API, we need to specify the size of the image we want to download. More info about this can be found here.

Let’s get back to our script. We need to select a date for retrieving interesting images on that date. Ideally, this date should be at least two days before. The limit for retrieving the number of images should be around five as you will need only one wallpaper daily. We will now generate our API URL and request it via HTTP request which will return us an XML response which we have to parse (Sample XML Response from the Flickr Interesting API can be found here). Parsing XML in Python is quite simple and can be done with ElementTree XML . After parsing the response we have to save all of its necessary attributes in independent variables and also replace any whitespaces with an underscore (to avoid any kind of conflicts while setting the wallpaper). With the help of these attributes we will generate our static URLs and add them in List[]. Also, we will create a Dictionary{} to maintain a mapping between image titles and URLs {'photo_url': 'photo_title'}.

So, we have five image URLs in a list and we can choose any one randomly (if you want) to set as our wallpaper. Once we have our desired image URL we can download the image file and save it in a directory. In this script, I have set the download directory as: ~/Pictures/Flickr/

If this download directory doesn’t exist it will be created so that images can be saved there (in .jpg format). This script is specifically targeted to Cinnamon Desktop Environment [Linux Mint 17.X]. Wallpapers can be changed in Cinnamon as follows:

gsettings set org.cinnamon.desktop.background picture-uri file:////absolute_path_of_image_no_spaces

The same can be achieved in Ubuntu with Unity DE as follows:

gsettings set org.gnome.desktop.background picture-uri file:////absolute_path_of_image_no_spaces

Now this script sets wallpaper only if there is no image for that day in the download directory, but you can change that logic according to your preferences. To automate this script ie. to execute it whenever you are connected to the internet you can try this software Cuttlefish.  More info about Cuttlefish at Ubuntu handbook and OMGUbuntu. If you want to do it the hard way refer the Ubuntu guide for OnNetworkConnectionRunScript and this post on AskUbuntu.

This slideshow requires JavaScript.

This script is available on Github you can modify it as per your needs. If you encounter any bugs let me now.

image4233

Dealing with compatibility issues.

ubuntu-logo14
Standard

Today I was searching for an Ubuntu app called Cuttlefish. But later found out that even though it was such an important utility it was no longer maintained. So, I googled ‘Cuttlefish ubuntu no longer maintained‘ and stumbled upon this AskUbuntu.com question…

How do I develop software for Ubuntu that one could still use in future versions

This question had a very thoughtful response, this point should be kept in mind while developing your projects…

In my experience (near 30 years now), hardware and low level code (BIOS for instance) change not that much or more accurately, keep a very good ascending compatibility. (I wrote twenty five years ago a little game program in assembler, running with DOS and VGA display: it’s still running fine on modern Windows computers).

So I would say: avoid dependencies as much as possible. When planning to use a library or an API, examine thoroughly its history and evolution, and how it still run “obsolete” code or not. If you’re in doubt, try to incorporate to your project the source code (and not the compiled library). If the functionalities it provides is not strictly computer oriented (like maths libs or general algorithms), you probably need not upgrades.

Just my two cents…

Find the whole story here : Ask Ubuntu

And as always:

Make it Open Source! That way anyone can jump in and maintain, fix, whatever, when for whatever reason you just can’t anymore, although this can happen due to basic operability (stares at air apps…) some apps just need an extra hit with a hammer to keep them working.