Search Results

Search found 8288 results on 332 pages for 'proxy models'.

Page 158/332 | < Previous Page | 154 155 156 157 158 159 160 161 162 163 164 165  | Next Page >

  • Following Domain Driven Design with MVVM/WPF

    - by msfanboy
    Hello, I have plain POCOs here and as INotifyPropertyChanged is a Interface for the Views need its implemented in the ViewModel not the Model. Now I want to show validation errors in the View beside every textbox the user typed in data. I do not want to implemented the IDataErrorInfo interface in my Models because lets assume I am not allowed to touch them as they come from another Service/Supplier. I do not want to put my IsCustomerFirstNameLenthValid Method into the Model because I could not have access to it or I just dont want to pollute my Models with interface`s having nothing to do there! How can I validate my naked POCO`s in the ViewModel and forward the results to the View by showing validation errors ?

    Read the article

  • How to avoid injecting dependencies into an object so that it can pass them on?

    - by Pheter
    I am interested in applying dependency injection to my current project, which makes use of the MVC pattern. My controllers will call the models and therefore will need to inject the dependencies into the models. To do this, the controller must have the dependencies (such as a database object) in the first place. The controller doesn't need to make use of some of these dependencies (such as the database object), so I feel that it shouldn't be given this dependency. However, it has to have these dependencies if it is to inject them into the model objects. How can I avoid having dependencies injected into an object just so that it can pass them on? Doing so feels wrong and can result in many dependencies being injected into an object. Edit: I am using PHP.

    Read the article

  • Is my form password being passed in clear text?

    - by liinkas
    This is what my browser sent, when logging into some site: POST http://www.some.site/login.php HTTP/1.0 User-Agent: Opera/8.26 (X2000; Linux i686; Z; en) Host: www.some.site Accept: text/html, application/xml;q=0.9, application/xhtml+xml, image/png, image/jpeg, image/gif, image/x-xbitmap, */*;q=0.1 Accept-Language: en-US,en;q=0.9 Accept-Charset: iso-8859-1, utf-8, utf-16, *;q=0.1 Accept-Encoding: deflate, gzip, x-gzip, identity, *;q=0 Referer: http://www.some.site/ Proxy-Connection: close Content-Length: 123 Content-Type: application/x-www-form-urlencoded lots_of_stuff=here&e2ad811=my_login_name&e327696=my_password&lots_of_stuff=here Can I state that anyone can sniff my login name and password for that site? Maybe just on my LAN? If so (even only on LAN ) then I'm shocked. I thought using <input type="password"> did something more than make all characters look like ' * ' p.s. If it matters I played with netcat (on linux) and made connection browser <= netcat (loged here) <= proxy <= remote_site

    Read the article

  • Why does sorl.thumbnail ImageField fail in the admin?

    - by Mark0978
    I have code that looks like this: from sorl.thumbnail import ImageField class Gallery(models.Model): pass class GalleryImage(models.Model): image = ImageField(upload_to='galleries') In the admin: class GalleryImageInline(admin.TabularInline): model = GalleryImage class GalleryAdmin(admin.ModelAdmin): inlines = (GalleryImageInline,) If I use the sorl.thumbnail as above, it is impossible to add images in the admin. I get the validation error Enter a list of values. If I replace the sorl.thumbnail.ImageField with a plain django ImageField, everything works. If I want sorl.thumbnail to clean up the cache thumbnails, I need to use it in the model, but if I use it in the model, I can't seem to add any images to need thumbnails. Anyone else found and fixed this problem yet?

    Read the article

  • Cant get the catch statement to catch the custom exception

    - by user282807
    i have created a custom exception in the business layer and also using wcf layer where I am calling the methods in the business layer then in another website i am calling the method from wcf. i can see the message that i wrote in custom exception but the program goes staright to exception (the second catch block) instead of hitting my first catch block(where the custom exception is) when i hover over the exception i see my message but it's inside something called faultexception which i am not familiar with. and in there under details..there i see type= CanOnlyApplyOnceException. here is my code: protected void AddNewApplication() { try { using (var proxy = new ServiceReference1.ServiceClient()) { proxy.AddApplication(new Application { Credentials = 2, Comments = txtComments.Text, }); } } catch (CanOnlyApplyOnceException c) { ErrorSummary.AddError(c.Message, this); return; } catch (Exception) { lblStatus.Text = "There has been an error. Please try again"; } }

    Read the article

  • Use only some parts of Django?

    - by Hanno Fietz
    I like Django, but for a particular application I would like to use only parts of it, but I'm not familiar enough with how Django works on the inside, so maybe someone can point me into the right direction as to what I have to check out. Specifically, I want to use: The models and database abstraction The caching API, although I want to avoid database lookups by caching, not HTML generation, and since the caching framework in Django is intended for the latter, I'm not sure yet whether that's really appropriate. I would not use: Templating urlconfigs Or, more exactly, I'm neither using HTTP nor HTML. So basically, I have a different input / output chain than usual. Can this work? My personal killer feature in Django is the Object / database mapping that I can do with the models, so if there's another technology (doesn't have to be Python, I'm in the design phase and I'm pretty agnostic about languages and platforms) that gives me the same abilities, that would be great, too.

    Read the article

  • graphviz dot: how to insert arrows from a node to center of an arrow

    - by gregor
    Hi everyone, I try to create diagrams for MPLUS analyses with dot from the graphviz package. Does anybody have experience with using dot to visualize structural equation models/latent class mixture models? There is especially one feature that I can't figure out how to do beautifully: I need arrows from nodes to the center of another arrow like C | | V A ------------> B I tried to insert an invisible node at the intersection of the arrows. This, however, results in a "cracked" A---B arrow because dot does represent it as two independent arrows. Is this even possible with dot? Thanks for suggestions and help! Gregor

    Read the article

  • Problems finding classes in namespace

    - by Matt
    I am trying to find all of the types in the Models namespace within an ASP.NET MVC assembly from within a testing assembly. I was trying to use LINQ to find the relevant set for me but it is returning an empty set on me. I am sure it is some simple mistake, I am still relatively new to LINQ admittedly. var abstractViewModelType = typeof (AbstractViewModel); var baseAssembly = Assembly.GetAssembly(abstractViewModelType); var modelTypes = baseAssembly.GetTypes() .Where(assemblyType => (assemblyType.Namespace.EndsWith("Models") && assemblyType.Name != "AbstractViewModel")) .Select(assemblyType => assemblyType); foreach(var modelType in modelTypes) { //Assert some things } When I reach the foreach I receive a Null reference exception.

    Read the article

  • Best practices for using memcached in Rails?

    - by Matt
    Hello everybody, as database transcations in our app are getting more and more time consuming, we have started to use memcached to reduce the amount of queries passed to MySQL. All in all, it works fine and really saves a lot of time. But as caching was "silently appearing" as a workaround to give the app more juice, a lot of our models now contain code like this: def self.all_cached Rails.cache.fetch('object_name') { find( :all, :include => [associations]) } end This is getting more and more a pain as filling and flushing the cache happens in several classes accross the application. Now, I was wondering if there was a better way to abstract memcached logic to make it more powerful and easy to use across all needed models? I was thinking about having some kind of memcached-module which is included in all needed modules. But before playing around, I thought: Let's ask experts first :-) Thanks Matt

    Read the article

  • CakePHP HABTM: Editing one item casuses HABTM row to get recreated, destroys extra data

    - by leo-the-manic
    I'm having trouble with my HABTM relationship in CakePHP. I have two models like so: Department HABTM Location. One large company has many buildings, and each building provides a limited number of services. Each building also has its own webpage, so in addition to the HABTM relationship itself, each HABTM row also has a url field where the user can visit to find additional information about the service they're interested and how it operates at the building they're interested in. I've set up the models like so: <?php class Location extends AppModel { var $name = 'Location'; var $hasAndBelongsToMany = array( 'Department' => array( 'with' => 'DepartmentsLocation', 'unique' => true ) ); } ?> <?php class Department extends AppModel { var $name = 'Department'; var $hasAndBelongsToMany = array( 'Location' => array( 'with' => 'DepartmentsLocation', 'unique' => true ) ); } ?> <?php class DepartmentsLocation extends AppModel { var $name = 'DepartmentsLocation'; var $belongsTo = array( 'Department', 'Location' ); // I'm pretty sure this method is unrelated. It's not being called when this error // occurs. Its purpose is to prevent having two HABTM rows with the same location // and department. function beforeSave() { // kill any existing rows with same associations $this->log(__FILE__ . ": killing existing HABTM rows", LOG_DEBUG); $result = $this->find('all', array("conditions" => array("location_id" => $this->data['DepartmentsLocation']['location_id'], "department_id" => $this->data['DepartmentsLocation']['department_id']))); foreach($result as $row) { $this->delete($row['DepartmentsLocation']['id']); } return true; } } ?> The controllers are completely uninteresting. The problem: If I edit the name of a Location, all of the DepartmentsLocations that were linked to that Location are re-created with empty URLs. Since the models specify that unique is true, this also causes all of the newer rows to overwrite the older rows, which essentially destroys all of the URLs. I would like to know two things: Can I stop this? If so, how? And, on a less technical and more whiney note: Why does this even happen? It seems bizarre to me that editing a field through Cake should cause so much trouble, when I can easily go through phpMyAdmin, edit the Location name there, and get exactly the result I would expect. Why does CakePHP touch the HABTM data when I'm just editing a field on a row? It's not even a foreign key!

    Read the article

  • .NET template class instance - passing a variable data type

    - by FerretallicA
    As the title suggests, I'm tyring to pass a variable data type to a template class. Something like this: frmExample = New LookupForm(Of Models.MyClass) 'Works fine Dim SelectedType As Type = InstanceOfMyClass.GetType() 'Works fine frmExample = New LookupForm(Of SelectedType) 'Ba-bow! frmExample = New LookupForm(Of InstanceOfMyClass.GetType()) 'Ba-bow! LookupForm<Models.MyClass> frmExample; Type SelectedType = InstanceOfMyClass.GetType(); frmExample = new LookupForm<SelectedType.GetType()>(); //Ba-bow frmExample = new LookupForm<(Type)SelectedType>(); //Ba-bow I'm assuming it's something to do with the template being processed at compile time but even if I'm off the mark there, it wouldn't solve my problem anyway. I can't find any relevant information on using Reflection to instance template classes either. (How) can I create an instance of a dynamically typed repository at runtime?

    Read the article

  • Consuming an RPC/encoded Web Service in .NET

    - by Timmy O' Tool
    I'm trying to build the proxy class for a web service using the wsdl Executing this command: wsdl [http://WSDL_URL] I'm getting Warning: This web reference does not conform to WS-I Basic Profile v1.1. R2706: A wsdl:binding in a DESCRIPTION MUST use the value of "literal" for the use attribute in all soapbind:body, soapbind:fault, soapbind:header and soapbind:headerfault elements. ... Error: Cannot find definition for http://schemas.xmlsoap.org/wsdl/:BouBinding. Service Description with namespace http://schemas.xmlsoap.org/wsdl/ is missing. Parameter name: name The author of the web service told me that the SOAP protocol is RPC/encoded. Is there is any way to generate a proxy class for this?

    Read the article

  • Is there a bash shortcut for traversing similar directory structures?

    - by Steve Weet
    The Korn shell used to have a very useful option to cd for traversing similar directory structures e.g. given the following directorys /home/sweet/dev/projects/trunk/projecta/app/models /home/andy/dev/projects/trunk/projecta/app/models Then if you were in the /home/sweet.... directory then you could change to the equivalent directory in andy's structure by typing cd sweet andy So if ksh saw 2 arguments then it would scan the current directory path for the first value, replace it with the second and cd there. Is anyone aware of similar functionality in bash. EDIT 1 Following on from Michal's excellent answer I have now created the following bash function called scd (For Sideways cd) function scd { cd "${PWD/$1/$2}" } EDIT 2 Thanks to @digitalross I can now reproduce the ksh functionality exactly with the code from below (With the addition of a pwd to tell you where you have changed to) cd () { if [ "x$2" != x ]; then builtin cd ${PWD/$1/$2} pwd else builtin cd "$@" fi }

    Read the article

  • iphone using 2 tableview

    - by Anonymous
    Hi, I 'm a iphone programming newbie. I m trying to implement an app with multi-level tableviews. The idea is if someone selects something on the first screen say then a new view opens(tableview) car make> list of models Honda > Acura, S2000,accord ... My challenge is to show the list of models in a new tableview & secondly change the size of hte list depending on the car selected. I have programmed so that when i select 'Honda' a new tableview opens. How do i populate data for the second table?? Any suggestions on how to proceed? thanks

    Read the article

  • Accessing constructor from abstract base class with reflection

    - by craesh
    Hi! I'm playing around with Java's Reflection. I have an abstract class Base with a constructor. abstract class Base { public Base( String foo ) { // do some magic } } I have some further classes extending Base. They don't contain much logic. I want to instantiate them with Base's constructor, without having to write some proxy contructors in those derived classes. And of course, I want to instantiate those derived classes with Reflection. Say: Class cls = SomeDerivedClass.class; Constructor constr; constr = cls.getConstructor( new Class[] { String.class } ); // will return null Class clsBase = Base.class; constr = clsBase.getConstructor( new Class[] { String.class } ); // ok Base obj = (Base) constr.newInstance( new Object[] { "foo" } ); // will throw InstantiationException because it belongs to an abstract class Any ideas, how I can instantiate a derived class with Base's constructor? Or must I declare those dumb proxy constructors?

    Read the article

  • MVVM Binding Orthogonal Aspects in Views e.g. Application Settings

    - by chibacity
    I have an application which I am developing using WPF\Prism\MVVM. All is going well and I have some pleasing MVVM implementations. However, in some of my views I would like to be able to bind application settings e.g. when a user reloads an application, the checkbox for auto-scrolling a grid should be checked in the state it was last time the user used the application. My view needs to bind to something that holds the "auto-scroll" setting state. I could put this on the view-model, but applications settings are orthogonal to the purpose of the view-model. The "auto-scroll" setting is controlling an aspect of the view. This setting is just an example. There will be quite a number of them and splattering my view-models with properties to represent application settings (so I can bind them) feels decidedly yucky. One view-model per view seems to be de rigeuer... What is best\usual practice here? Splatter my view-models with application settings? Have multiple view-models per view so settings can be represented in their own right? Split views so that controls can bind to an ApplicationSettingsViewModel? = too many views? Something else? Edit 1 To add a little more context, I am developing a UI with a tabbed interface. Each tab will host a single widget and there a variety of widgets. Each widget is a Prism composition of individual views. Some views are common amongst widgets e.g. a file picker view. Whilst each widget is composed of several views, as a whole, conceptually a widget has a single set of user settings e.g. last file selected, auto-scroll enabled, etc. These need to be persisted and retrieved\applied when the application starts again, and the widget views are created. My question is focused on the fact that conceptually a widget has a single set of user settings which is at right-angles to the fact that a widget consists of many views. Each view in the widget has it's own view-model (which works nicely and logically) but if I stick to a one view-model per view, I would have to splatter each view-model with user settings appropriate to it. This doesn't sound right ?!?

    Read the article

  • Graphics glitch when drawing to a Cairo context obtained from a gtk.DrawingArea inside a gtk.Viewport.

    - by user410023
    I am trying to redraw the part of the DrawingArea that is visible in the Viewport in the expose-event handler. However, it seems that I am doing something wrong with the coordinates that are passed to the event handler because there is garbage at the edge of the Viewport when scrolling. Can anyone tell what I am doing wrong? Here is a small example: import pygtk pygtk.require("2.0") import gtk from numpy import array from math import pi class Circle(object): def init(self, position = [0., 0.], radius = 0., edge = (0., 0., 0.), fill = None): self.position = position self.radius = radius self.edge = edge self.fill = fill def draw(self, ctx): rect = array(ctx.clip_extents()) rect[2] -= rect[0] rect[3] -= rect[1] center = rect[2:4] / 2 ctx.arc(center[0], center[1], self.radius, 0., 2. * pi) if self.fill != None: ctx.set_source_rgb(*self.fill) ctx.fill_preserve() ctx.set_source_rgb(*self.edge) ctx.stroke() class Scene(object): class Proxy(object): directory = {} def init(self, target, layers = set()): self.target = target self.layers = layers Scene.Proxy.directory[target] = self def __init__(self, viewport): self.objects = {} self.layers = [set()] self.viewport = viewport self.signals = {} def draw(self, ctx): x = self.viewport.get_hadjustment().value y = self.viewport.get_vadjustment().value ctx.set_source_rgb(1., 1., 1.) ctx.paint() ctx.translate(x, y) for obj in self: obj.draw(ctx) def add(self, item, layer = 0): item = Scene.Proxy(item, layers = set((layer,))) assert(hasattr(item.target, "draw")) assert(isinstance(layer, int)) item.layers.add(layer) while not layer < len(self.layers): self.layers.append(set()) self.layers[layer].add(item) if not item in self.objects: self.objects[item] = set() self.objects[item].add(layer) def remove(self, item, layers = None): item = Scene.Proxy.directory[item] if layers == None: layers = self.objects[item] for layer in layers: layer.remove(item) item.layers.remove(layer) if len(item.layers) == 0: self.objects.remove(item) def __iter__(self): for layer in self.layers: for item in layer: yield item.target class App(object): def init(self): signals = { "canvas_exposed": self.update_canvas, "gtk_main_quit": gtk.main_quit } self.builder = gtk.Builder() self.builder.add_from_file("graphics_glitch.glade") self.window = self.builder.get_object("window") self.viewport = self.builder.get_object("viewport") self.canvas = self.builder.get_object("canvas") self.scene = Scene(self.viewport) signals.update(self.scene.signals) self.builder.connect_signals(signals) self.window.show() def update_canvas(self, widget, event): ctx = self.canvas.window.cairo_create() self.scene.draw(ctx) ctx.clip() if name == "main": app = App() scene = app.scene scene.add(Circle((0., 0.), 10.)) gtk.main() And the Glade file "graphics_glitch.glade": <?xml version="1.0"?> <interface> <requires lib="gtk+" version="2.16"/> <!-- interface-naming-policy project-wide --> <object class="GtkWindow" id="window"> <property name="width_request">200</property> <property name="height_request">200</property> <property name="visible">True</property> <signal name="destroy" handler="gtk_main_quit"/> <child> <object class="GtkScrolledWindow" id="scrolledwindow1"> <property name="visible">True</property> <property name="can_focus">True</property> <property name="hadjustment">h_adjust</property> <property name="vadjustment">v_adjust</property> <property name="hscrollbar_policy">automatic</property> <property name="vscrollbar_policy">automatic</property> <child> <object class="GtkViewport" id="viewport"> <property name="visible">True</property> <property name="resize_mode">queue</property> <child> <object class="GtkDrawingArea" id="canvas"> <property name="width_request">640</property> <property name="height_request">480</property> <property name="visible">True</property> <signal name="expose_event" handler="canvas_exposed"/> </object> </child> </object> </child> </object> </child> </object> <object class="GtkAdjustment" id="h_adjust"> <property name="lower">-1000</property> <property name="upper">1000</property> <property name="step_increment">1</property> <property name="page_increment">25</property> <property name="page_size">25</property> </object> <object class="GtkAdjustment" id="v_adjust"> <property name="lower">-1000</property> <property name="upper">1000</property> <property name="step_increment">1</property> <property name="page_increment">25</property> <property name="page_size">25</property> </object> </interface> Thanks! --Dan

    Read the article

  • HOWTO: Deserialize WCF message using OperationContract

    - by Stefan
    Hi, I succeeded in building a WCF client generated by svcutil.exe from the WSDL. Using the generated client proxy class I can call the web service of an external service supplier. I also succeeded in coding a message inspector, as I need to log both raw XML request and response as full SOAP message to the database. For an emergency szenario I also need to be able to "import" a raw XML response. I found many hints on using XMLSerializer or deserializing WCF messages based on the message contract. But how can I deserialize a raw XML response based on an operation contract? For a first test I use one of the logged raw responses, save it to a file and now try to deserialize it to the response type as generated in the client proxy. Somehow I must succeed in calling DeserializeReply() from class ClientOperation. But how to get there? I happily acced any help as I'm quite new to WCF... TIA, Stefan

    Read the article

  • Observing an NSMutableArray for insertion/removal

    - by Adam Ernst
    A class has a property (and instance var) of type NSMutableArray with synthesized accessors (via @property). If you observe this array using: [myObj addObserver:self forKeyPath:@"theArray" options:0 context:NULL]; And then insert an object in the array like this: [[myObj theArray] addObject:[NSString string]]; An observeValueForKeyPath... notification is not sent. However, the following does send the proper notification: [[myObj mutableArrayValueForKey:@"theArray"] addObject:[NSString string]]; This is because mutableArrayValueForKey returns a proxy object that takes care of notifying observers. But shouldn't the synthesized accessors automatically return such a proxy object? What's the proper way to work around this--should I write a custom accessor that just invokes [super mutableArrayValueForKey...]?

    Read the article

  • Looking for Fiddler2 help. connection to gateway refused? Just got rid of a virus

    - by John Mackey
    I use Fiddler2 for facebook game items, and it's been a great success. I accessed a website to download some dat files I needed. I think it was eshare, ziddu or megaupload, one of those. Anyway, even before the rar file had downloaded, I got this weird green shield in the bottom right hand corner of my computer. It said a Trojan was trying to access my computer, or something to that extent. It prompted me to click the shield to begin anti-virus scanning. It turns out this rogue program is called Antivirus System Pro and is pretty hard to get rid of. After discovering the rogue program, I tried using Fiddler and got the following error: [Fiddler] Connection to Gateway failed.Exception Text: No connection could be made because the target machine actively refused it 127.0.0.1:5555 I ended up purchasing SpyDoctor + Antivirus, which I'm told is designed specifically for getting rid of these types of programs. Anyway, I did a quick-scan last night with spydoctor and malware bytes. Malware picked up 2 files, and Spydoctor found 4. Most were insignificant, but it did find a worm called Worm.Alcra.F, which was labeled high-priority. I don’t know if that’s the Anti-Virus Pro or not, but SpyDoctor said it got rid of all of those successfully. I tried to run Fiddler again before leaving home, but was still getting the "gateway failed" error. Im using the newest version of firefox. When I initially set up the Fiddler 2.2.8.6, I couldn’t get it to run at first, so I found this faq on the internet that said I needed to go through ToolsOptionsSettings and set up an HTTP Proxy to 127.0.0.1 and my Port to 8888. Once I set that up and downloaded this fiddler helper as a firefox add-on, it worked fine. When I turn on fiddler, it automatically takes my proxy setting from no proxy (default) to the 127.0.0.1 with Port 8888 set up. It worked fine until my computer detected this virus. Anyway, hopefully I've given you sufficient information to offer me your best advice here. Like I said, Spydoctor says the bad stuff is gone, so maybe the rogue program made some type of change in my fiddler that I could just reset or uncheck or something like that? Or will I need to completely remove fiddler and those dat files and rar files I downloaded? Any help would be greatly appreciated. Thanks for your time.

    Read the article

  • Cant access websites [closed]

    - by LiveEn
    Recently i have some problems accessing websites. When i try to access it says This webpage is not available. I tried accessing the site through FireFox, Internet Explorer and Chrome. I also tried with using a web proxy but still the same problem. This problem is only in my Desktop PC, all the websites works fine in my laptop Currently i cant access, yahoo.com download.com bing.com proxy.org daniweb.com aol.com and many forums I check with the host file but nothing is blocked in that. Can some one please suggest what is wrong?? Thanks

    Read the article

  • ACL suggestions needed for a small light framework developed

    - by Sai
    Hello all, I've developed a small framework that was needed for the firm that I work. The problem is I've not made a full level framework, as developers are finding tough time to understand what is what, So I made a simple structure with app/controllers, app/models, app/views so that they can clearly separate the code and can get used to higher frameworks later. All other things like components/modules/helpers are just coded for now in a file called app_functions.php. The above framework is derived from cakePHP, not all, but just the framework uses a modified code of dispatcher of cake. The models all work with normal sqls, and views with pure html using a template engine. Now I'm trying to develop a small application over it, it cameout well, but ACL is something which many versions of it have not proved that satisfactory. Can anybody suggest a good ACL. I've checked phpGACL(too heavy), Zend_ACL likely won't be compatible, Cakephp's ACL which wont fit in the structure,as we are not using cake's SQLs, as there too many queries hitting the database for one page display.

    Read the article

  • Dealing with dependencies between WCF services when using Castle Windsor

    - by Georgia Brown
    I have several WCF services which use castle windsor to resolve their dependencies. Now I need some of these services to talk to each other. The typical structure is service -- Business Logic -- DAL The calls to the other services need to occur at Business Logic level. What is the best approach for implementing this? Should I simply inject a service proxy into the business logic? Is this wasteful if for example, only one of two method from my service need to use this proxy? What if the services need to talk to each other? - Will castle windsor get stuck in a loop trying to resolve each services dependencies?

    Read the article

  • Ngram IDF smoothing

    - by adi92
    I am trying to use IDF scores to find interesting phrases in my pretty huge corpus of documents. I basically need something like Amazon's Statistically Improbable Phrases, i.e. phrases that distinguish a document from all the others The problem that I am running into is that some (3,4)-grams in my data which have super-high idf actually consist of component unigrams and bigrams which have really low idf.. For example, "you've never tried" has a very high idf, while each of the component unigrams have very low idf.. I need to come up with a function that can take in document frequencies of an n-gram and all its component (n-k)-grams and return a more meaningful measure of how much this phrase will distinguish the parent document from the rest. If I were dealing with probabilities, I would try interpolation or backoff models.. I am not sure what assumptions/intuitions those models leverage to perform well, and so how well they would do for IDF scores. Anybody has any better ideas?

    Read the article

< Previous Page | 154 155 156 157 158 159 160 161 162 163 164 165  | Next Page >