Search Results

Search found 1348 results on 54 pages for 'floating accuracy'.

Page 22/54 | < Previous Page | 18 19 20 21 22 23 24 25 26 27 28 29  | Next Page >

  • How do NTP Servers Manage to Stay so Accurate?

    - by Akemi Iwaya
    Many of us have had the occasional problem with our computers and other devices retaining accurate time settings, but a quick sync with an NTP server makes all well again. But if our own devices can lose accuracy, how do NTP servers manage to stay so accurate? Today’s Question & Answer session comes to us courtesy of SuperUser—a subdivision of Stack Exchange, a community-driven grouping of Q&A web sites. Photo courtesy of LEOL30 (Flickr). The Question SuperUser reader Frank Thornton wants to know how NTP servers are able to remain so accurate: I have noticed that on my servers and other machines, the clocks always drift so that they have to sync up to remain accurate. How do the NTP server clocks keep from drifting and always remain so accurate? How do the NTP servers manage to remain so accurate? The Answer SuperUser contributor Michael Kjorling has the answer for us: NTP servers rely on highly accurate clocks for precision timekeeping. A common time source for central NTP servers are atomic clocks, or GPS receivers (remember that GPS satellites have atomic clocks onboard). These clocks are defined as accurate since they provide a highly exact time reference. There is nothing magical about GPS or atomic clocks that make them tell you exactly what time it is. Because of how atomic clocks work, they are simply very good at, having once been told what time it is, keeping accurate time (since the second is defined in terms of atomic effects). In fact, it is worth noting that GPS time is distinct from the UTC that we are more used to seeing. These atomic clocks are in turn synchronized against International Atomic Time or TAI in order to not only accurately tell the passage of time, but also the time. Once you have an exact time on one system connected to a network like the Internet, it is a matter of protocol engineering enabling transfer of precise times between hosts over an unreliable network. In this regard a Stratum 2 (or farther from the actual time source) NTP server is no different from your desktop system syncing against a set of NTP servers. By the time you have a few accurate times (as obtained from NTP servers or elsewhere) and know the rate of advancement of your local clock (which is easy to determine), you can calculate your local clock’s drift rate relative to the “believed accurate” passage of time. Once locked in, this value can then be used to continuously adjust the local clock to make it report values very close to the accurate passage of time, even if the local real-time clock itself is highly inaccurate. As long as your local clock is not highly erratic, this should allow keeping accurate time for some time even if your upstream time source becomes unavailable for any reason. Some NTP client implementations (probably most ntpd daemon or system service implementations) do this, and others (like ntpd’s companion ntpdate which simply sets the clock once) do not. This is commonly referred to as a drift file because it persistently stores a measure of clock drift, but strictly speaking it does not have to be stored as a specific file on disk. In NTP, Stratum 0 is by definition an accurate time source. Stratum 1 is a system that uses a Stratum 0 time source as its time source (and is thus slightly less accurate than the Stratum 0 time source). Stratum 2 again is slightly less accurate than Stratum 1 because it is syncing its time against the Stratum 1 source and so on. In practice, this loss of accuracy is so small that it is completely negligible in all but the most extreme of cases. Have something to add to the explanation? Sound off in the comments. Want to read more answers from other tech-savvy Stack Exchange users? Check out the full discussion thread here.

    Read the article

  • To sample or not to sample...

    - by [email protected]
    Ideally, we would know the exact answer to every question. How many people support presidential candidate A vs. B? How many people suffer from H1N1 in a given state? Does this batch of manufactured widgets have any defective parts? Knowing exact answers is expensive in terms of time and money and, in most cases, is impractical if not impossible. Consider asking every person in a region for their candidate preference, testing every person with flu symptoms for H1N1 (assuming every person reported when they had flu symptoms), or destructively testing widgets to determine if they are "good" (leaving no product to sell). Knowing exact answers, fortunately, isn't necessary or even useful in many situations. Understanding the direction of a trend or statistically significant results may be sufficient to answer the underlying question: who is likely to win the election, have we likely reached a critical threshold for flu, or is this batch of widgets good enough to ship? Statistics help us to answer these questions with a certain degree of confidence. This focuses on how we collect data. In data mining, we focus on the use of data, that is data that has already been collected. In some cases, we may have all the data (all purchases made by all customers), in others the data may have been collected using sampling (voters, their demographics and candidate choice). Building data mining models on all of your data can be expensive in terms of time and hardware resources. Consider a company with 40 million customers. Do we need to mine all 40 million customers to get useful data mining models? The quality of models built on all data may be no better than models built on a relatively small sample. Determining how much is a reasonable amount of data involves experimentation. When starting the model building process on large datasets, it is often more efficient to begin with a small sample, perhaps 1000 - 10,000 cases (records) depending on the algorithm, source data, and hardware. This allows you to see quickly what issues might arise with choice of algorithm, algorithm settings, data quality, and need for further data preparation. Instead of waiting for a model on a large dataset to build only to find that the results don't meet expectations, once you are satisfied with the results on the initial sample, you can  take a larger sample to see if model quality improves, and to get a sense of how the algorithm scales to the particular dataset. If model accuracy or quality continues to improve, consider increasing the sample size. Sampling in data mining is also used to produce a held-aside or test dataset for assessing classification and regression model accuracy. Here, we reserve some of the build data (data that includes known target values) to be used for an honest estimate of model error using data the model has not seen before. This sampling transformation is often called a split because the build data is split into two randomly selected sets, often with 60% of the records being used for model building and 40% for testing. Sampling must be performed with care, as it can adversely affect model quality and usability. Even a truly random sample doesn't guarantee that all values are represented in a given attribute. This is particularly troublesome when the attribute with omitted values is the target. A predictive model that has not seen any examples for a particular target value can never predict that target value! For other attributes, values may consist of a single value (a constant attribute) or all unique values (an identifier attribute), each of which may be excluded during mining. Values from categorical predictor attributes that didn't appear in the training data are not used when testing or scoring datasets. In subsequent posts, we'll talk about three sampling techniques using Oracle Database: simple random sampling without replacement, stratified sampling, and simple random sampling with replacement.

    Read the article

  • Countdown of Top 10 Reasons to Never Ever Use a Pie Chart

    - by Tony Wolfram
      Pie charts are evil. They represent much of what is wrong with the poor design of many websites and software applications. They're also innefective, misleading, and innacurate. Using a pie chart as your graph of choice to visually display important statistics and information demonstrates either a lack of knowledge, laziness, or poor design skills. Figure 1: A floating, tilted, 3D pie chart with shadow trying (poorly)to show usage statistics within a graphics application.   Of course, pie charts in and of themselves are not evil. This blog is really about designers making poor decisions for all the wrong reasons. In order for a pie chart to appear on a web page, somebody chose it over the other alternatives, and probably thought they were doing the right thing. They weren't. Using a pie chart is almost always a bad design decision. Figure 2: Pie Chart from an Oracle Reports User Guide   A pie chart does not do the job of effectively displaying information in an elegant visual form.  Being circular, they use up too much space while not allowing their labels to line up. Bar charts, line charts, and tables do a much better job. Expert designers, statisticians, and business analysts have documented their many failings, and strongly urge software and report designers not to use them. It's obvious to them that the pie chart has too many inherent defects to ever be used effectively. Figure 3: Demonstration of how comparing data between multiple pie charts is difficult.   Yet pie charts are still used frequently in today's software applications, financial reports, and websites, often on the opening page as a symbol of how the data inside is represented. In an attempt to get a flashy colorful graphic to break up boring text, designers will often settle for a pie chart that looks like pac man, a colored spinning wheel, or a 3D floating alien space ship.     Figure 4: Best use of a pie chart I've found yet.   Why is the pie chart so popular? Through its constant use and iconic representation as the classic chart, the idea persists that it must be a good choice, since everyone else is still using it. Like a virus or an urban legend, no amount of vaccine or debunking will slow down the use of pie charts, which seem to be resistant to logic and common sense. Even the new iPad from Apple showcases the pie chart as one of its options.     Figure 5: Screen shot of new iPad showcasing pie charts. Regardless of the futility in trying to rid the planet of this often used poor design choice, I now present to you my top 10 reasons why you should never, ever user a pie chart again.    Number 10 - Pie Charts Just Don't Work When Comparing Data Number 9 - You Have A Better Option: The Sorted Horizontal Bar Chart Number 8 - The Pie Chart is Always Round Number 7 - Some Genius Will Make It 3D Number 6 - Legends and Labels are Hard to Align and Read Number 5 - Nobody Has Ever Made a Critical Decision Using a Pie Chart Number 4 - It Doesn't Scale Well to More Than 2 Items Number 3 - A Pie Chart Causes Distortions and Errors Number 2 - Everyone Else Uses Them: Debunking the "Urban Legend" of Pie Charts Number 1 - Pie Charts Make You Look Stupid and Lazy  

    Read the article

  • Cutting Subscriber Churn with Media Intelligence

    - by Oracle M&E
    There's lots of talk in media and entertainment companies about using "big data".  But it's often hard to see through the hype and understand how big data brings benefits in the real world.  How about being able to predict with 92% accuracy which subscribers intend to cancel their subscription - and put in place a renewal strategy to dramatically reduce that churn?  That's what Belgian media company De Persgroep has achieved with Oracle's Media Intelligence solution.  "One of the areas in which we're able to achieve beautiful results using big data is the churn prediction," De Persgroep's CIO Luc Verbist explains in a new Oracle video.  "Based on all the data that we collect on websites and all your behavior, payment behavior and so on, we're able to make a prediction model, which, with an accuracy of 92 percent, is able to predict that you probably won't renew your newspaper, anymore. So our approach to renewal is completely different to the people in that segment than towards the other people. And this has brought us a lot of value and a lot of customers who didn't stop their newspaper where else they would have done so." De Persgroep is using Oracle's Big Data Appliance, along with software from Oracle partner NGDATA to build up a detailed "DNA profile" of each individual customer, based on every interaction, in real time.  This means that any change in behavior - a drop in content consumption, a late subscription payment, a negative social media comment - is captured.  Applying advanced data modeling techniques automatically converts those raw interactions into data with real business meaning - like that customer's risk of churning. The very same data profile - comprising hundreds if individual dimensions - can simultaneously drive targeted marketing campaigns - informing audience about new content that's most relevant and encouraging them to subscribe.  It can power content recommendations and personalization right in the content sites and apps. And it can link directly into digital advertising networks via platforms like Oracle's BlueKai data management platform (DMP), to drive increased advertising CPMs. Using Oracle's Media Intelligence solution enables this across De Persgroep's business - comprising eight newspapers and 25 magazines published in Belgium and The Netherlands, and digital properties including websites with 6m daily unique visitors, along with TV and radio stations. "The company strategy is in fact a customer-centric strategy, so we want to get a 360-view about our customers, about our prospects. And the big data project helped us to achieve that goal," says Verbist. Using Oracle's Big Data Appliance to underpin the solution created huge savings.   "The selection of the Big Data Appliance was quite easy.  It was very quick to install, very easy to install, as well. And it was far cheaper than building our own Hadoop cluster. So it was in fact a non-brainer," Verbist explains. Applying Media Intelligence approach has yielded incredible results for De Persgroep, including: Improved products - with a new understanding of how readers are consuming print and digital content across the day Improved customer segmentation - driving a 6X improvement in customer prospecting and acquisition when contacting a specific segment Having the project up and running in three months And that has led to competitive benefits for De Persgroep, as Luc Verbist explains: "one of the results we saw since we started using big data is that we're able to increase the gap between we as the market leader, and the second [by] more than 20 percent."

    Read the article

  • Removing Barriers to Create Effective Data Models

    After years of creating and maintaining data models, I have started to notice common barriers that decrease the accuracy and usefulness of models. In my opinion, the main causes of these barriers are the lack of knowledge and communication from within a company. The lack of knowledge in regards to data models or data modeling can take many forms. Company Culture Knowledge Whether documented or undocumented, existing business rules of a company can affect how data is modeled. For example, if a company only allows 1 assigned person per customer to be able to manipulate a customer’s record then then a data model that includes an associated table that joins customers and employee’s would be unneeded because that would allow for the possibility of multiple employees to handle a customer because of the potential for a many to many relationship between Customers and Employees. Technical Knowledge Depending on the data modeler’s proficiency in modeling data they can inadvertently cause issues and/or complications with a design without even noticing. It is important that companies share data modeling responsibilities so that the models are developed from multiple perspectives of a system, company and the original problem.  In addition, the tools that a company selects to create data models can also affect the accuracy of the model if designer are not familiar with the tools or the tools are too complex to use for the designer. Existing System Knowledge In order for a data modeler to model data for an existing system so that new changes can be applied to a system then they need to at least know the basic concepts of a system so that they can work within it. This will promote reusability of data and prevent the chance of duplicating data. Project Knowledge This should be pretty obvious, but it is very hard to create an accurate data model without knowing what data needs to be modeled. I have always found it strange that I have been asked to start modeling data prior to a client formalizing any requirements. Usually when this happens I have to make several iterations to a model, and the client still does not know exactly what they want.  In addition additional issues can arise when certain stakeholders of a project are not consulted prior to the design or after the project is over because it can cause miss understandings and confusion by the end user as well as possibly not solving the original problem for which a project is intended to solve. One common thread between each type of knowledge is that they can all be avoided through the use of good communication. For example, if a modeler is new to a company then they should ask older employees about any business specific rules that may be documented or undocumented that must be applied to projects in general. Furthermore, if a modeler is not really familiar with a specific data modeling software then they need to speak up and ask for help form other employees or their manager. This will not only help the modeler in the project, but also help them in future projects that they do for the company. Additionally, if a project is not clearly defined prior to a data modeler being assigned the modeling project then it is their responsibility to communicate with the other stakeholders to clarify any part of a project that is unclear so that the data model that is created is accurately aligned with a project.

    Read the article

  • Where Have All the Ugly Forms Gone? Users and ADF Took Care Of It

    - by ultan o'broin
    Sometimes I hear that our application demos are a bit too "cutsey" and that we never talk about with any user roles that have lots of data entry as a requirement. Some (no names) consider those old clunker forms, with the myriad rows of fields, to be super-productive for data clerks. We do have such roles covered in Oracle Fusion Applications for sure. But consider what is really the issue here: productivity. Check out how the Oracle Fusion Financials Applications User Experience team went about designing for productivity when receiving and entering invoice data, for example. See how Fusion Financials caters so well for input and control of data? Central to all this is knowing the users and how they work: what tasks do they need to perform, and when. Read more about Fusion Financials productivity in the white paper, Get It Done Fast, Get It Done Right: The Oracle Fusion Financials User Experience. Now and then, I see forms that weren't designed for end user activity at all. Instead, they were designed by developers or by the IT department around the database schema. Forms with literally dozens of fields on the same page, sometimes. Forms that give the impression there was only task involved, when there may have been several. At times, completing one of these huge forms accurately became so tedious that, under pressure, it made more sense for the user to complete it quickly as possible and then let somebody else check it for accuracy and fill in the gaps from data emailed along in spreadsheet form. Data accuracy is critical in our business. Not good. Not efficient. Not productive. So here are a few basics on forms design for data entry-type user roles. A great place for developers to start exploring what is possible with forms layout is the Rich Client User Experience (RCUX) guidance on Form Layout, using ADF components. User-Centered Forms Design Considerations The starting point--something you must always keep in mind with your own design--is design for the end user. Find a representative end user, and keep that user engaged throughout the design, deployment, and test process. Consider these points in user testing those forms: Are there automated or technical solutions to entering the data that avoid manual input in the first place? For example, imports, uploads, OCR, whatever. Some day we will be able to tell Siri to do it, but leave that for now. Design your form to reflect the task involved (i.e., the business process) and not the database schema. On the form, group like fields together, logically. Eliminate duplicate data entry or prepopulate from previous data entry. Allow users to complete fields in the order they wish (i.e., no interdependency). Allow for tabbing between fields (keyboard is faster than mouse), so know how the browser supports this (see that RCUX guideline). Allow for final validation at the page level not at field-level entry. Way better for heads-down users. For example, ADF messages allow you to see a list of all validation errors on a page on a final submit or navigation action and to easily navigate to the point of error. Better still, be error tolerant. Allow users to enter data in formats they comfortable with. Bind any relevant user preference setting to the input format allowed (for example, the locale date format). Explore what data entry conversion can do for you automatically too (see the ADF converter demos, convenience patterns can also be written). Only ask for data input when it's needed. Get rid of, or hide optional fields. Cut down on the number of mandatory fields, and mark them clearly (use a *). Clearly label the fields in plain language. I am sure you may have a few more tips on forms design for data entry users. Remember the user before finding the comments.

    Read the article

  • Numerically stable(ish) method of getting Y-intercept of mouse position?

    - by Fraser
    I'm trying to unproject the mouse position to get the position on the X-Z plane of a ray cast from the mouse. The camera is fully controllable by the user. Right now, the algorithm I'm using is... Unproject the mouse into the camera to get the ray: Vector3 p1 = Vector3.Unproject(new Vector3(x, y, 0), 0, 0, width, height, nearPlane, farPlane, viewProj; Vector3 p2 = Vector3.Unproject(new Vector3(x, y, 1), 0, 0, width, height, nearPlane, farPlane, viewProj); Vector3 dir = p2 - p1; dir.Normalize(); Ray ray = Ray(p1, dir); Then get the Y-intercept by using algebra: float t = -ray.Position.Y / ray.Direction.Y; Vector3 p = ray.Position + t * ray.Direction; The problem is that the projected position is "jumpy". As I make small adjustments to the mouse position, the projected point moves in strange ways. For example, if I move the mouse one pixel up, it will sometimes move the projected position down, but when I move it a second pixel, the project position will jump back to the mouse's location. The projected location is always close to where it should be, but it does not smoothly follow a moving mouse. The problem intensifies as I zoom the camera out. I believe the problem is caused by numeric instability. I can make minor improvements to this by doing some computations at double precision, and possibly abusing the fact that floating point calculations are done at 80-bit precision on x86, however before I start micro-optimizing this and getting deep into how the CLR handles floating point, I was wondering if there's an algorithmic change I can do to improve this? EDIT: A little snooping around in .NET Reflector on SlimDX.dll: public static Vector3 Unproject(Vector3 vector, float x, float y, float width, float height, float minZ, float maxZ, Matrix worldViewProjection) { Vector3 coordinate = new Vector3(); Matrix result = new Matrix(); Matrix.Invert(ref worldViewProjection, out result); coordinate.X = (float) ((((vector.X - x) / ((double) width)) * 2.0) - 1.0); coordinate.Y = (float) -((((vector.Y - y) / ((double) height)) * 2.0) - 1.0); coordinate.Z = (vector.Z - minZ) / (maxZ - minZ); TransformCoordinate(ref coordinate, ref result, out coordinate); return coordinate; } // ... public static void TransformCoordinate(ref Vector3 coordinate, ref Matrix transformation, out Vector3 result) { Vector3 vector; Vector4 vector2 = new Vector4 { X = (((coordinate.Y * transformation.M21) + (coordinate.X * transformation.M11)) + (coordinate.Z * transformation.M31)) + transformation.M41, Y = (((coordinate.Y * transformation.M22) + (coordinate.X * transformation.M12)) + (coordinate.Z * transformation.M32)) + transformation.M42, Z = (((coordinate.Y * transformation.M23) + (coordinate.X * transformation.M13)) + (coordinate.Z * transformation.M33)) + transformation.M43 }; float num = (float) (1.0 / ((((transformation.M24 * coordinate.Y) + (transformation.M14 * coordinate.X)) + (coordinate.Z * transformation.M34)) + transformation.M44)); vector2.W = num; vector.X = vector2.X * num; vector.Y = vector2.Y * num; vector.Z = vector2.Z * num; result = vector; } ...which seems to be a pretty standard method of unprojecting a point from a projection matrix, however this serves to introduce another point of possible instability. Still, I'd like to stick with the SlimDX Unproject routine rather than writing my own unless it's really necessary.

    Read the article

  • Copy protection and licensing tools.

    - by Skittles
    I'm new to stackoverflow.com after hearing about it from Jon Skeet on DotNetRocks.This seems like the perfect place to ask this question. I am in the middle of trying to find a 3rd party Copy protection and licensing tool. The company that I work with have 4 products that need to be protected. We want to supply a Trail license (with extensions). A single user license and a floating license (where the client purchases a number to run over a network). We also want to be able to supply both the Single and Floating license as a subscription license. I have trialled DeployLX and although it seems to give everything that we need, and they are quick to answer emails, their documentation is truly awful with NO examples of how to achieve results. Has anyone any experience with DeployLX and if so, would you recommend it? Could you point me in the direction to find some real help on it? Finally, would anyone have any recommendations of a 3rd party licensing tool to use for very quick development. Thank you so much,

    Read the article

  • Eigenvector computation using OpenCV

    - by Andriyev
    Hi I have this matrix A, representing similarities of pixel intensities of an image. For example: Consider a 10 x 10 image. Matrix A in this case would be of dimension 100 x 100, and element A(i,j) would have a value in the range 0 to 1, representing the similarity of pixel i to j in terms of intensity. I am using OpenCV for image processing and the development environment is C on Linux. Objective is to compute the Eigenvectors of matrix A and I have used the following approach: static CvMat mat, *eigenVec, *eigenVal; static double A[100][100]={}, Ain1D[10000]={}; int cnt=0; //Converting matrix A into a one dimensional array //Reason: That is how cvMat requires it for(i = 0;i < affnDim;i++){ for(j = 0;j < affnDim;j++){ Ain1D[cnt++] = A[i][j]; } } mat = cvMat(100, 100, CV_32FC1, Ain1D); cvEigenVV(&mat, eigenVec, eigenVal, 1e-300); for(i=0;i < 100;i++){ val1 = cvmGet(eigenVal,i,0); //Fetching Eigen Value for(j=0;j < 100;j++){ matX[i][j] = cvmGet(eigenVec,i,j); //Fetching each component of Eigenvector i } } Problem: After execution I get nearly all components of all the Eigenvectors to be zero. I tried different images and also tried populating A with random values between 0 and 1, but the same result. Few of the top eigenvalues returned look like the following: 9805401476911479666115491135488.000000 -9805401476911479666115491135488.000000 -89222871725331592641813413888.000000 89222862280598626902522986496.000000 5255391142666987110400.000000 I am now thinking on the lines of using cvSVD() which performs singular value decomposition of real floating-point matrix and might yield me the eigenvectors. But before that I thought of asking it here. Is there anything absurd in my current approach? Am I using the right API i.e. cvEigenVV() for the right input matrix (my matrix A is a floating point matrix)? cheers

    Read the article

  • Code Golf: Leibniz formula for Pi

    - by Greg Beech
    I recently posted one of my favourite interview whiteboard coding questions in "What's your more controversial programming opinion", which is to write a function that computes Pi using the Leibniz formula. It can be approached in a number of different ways, and the exit condition takes a bit of thought, so I thought it might make an interesting code golf question. Shortest code wins! Given that Pi can be estimated using the function 4 * (1 - 1/3 + 1/5 - 1/7 + ...) with more terms giving greater accuracy, write a function that calculates Pi to within 0.00001. Edit: 3 Jan 2008 As suggested in the comments I changed the exit condition to be within 0.00001 as that's what I really meant (an accuracy 5 decimal places is much harder due to rounding and so I wouldn't want to ask that in an interview, whereas within 0.00001 is an easier to understand and implement exit condition). Also, to answer the comments, I guess my intention was that the solution should compute the number of iterations, or check when it had done enough, but there's nothing to prevent you from pre-computing the number of iterations and using that number. I really asked the question out of interest to see what people would come up with.

    Read the article

  • Milliseconds in DateTime.Now on .NET Compact Framework always zero? [SOLVED]

    - by Marcel
    Hi all, i want to have a time stamp for logs on a Windows Mobile project. The accuracy must be in the range a hundred milliseconds at least. However my call to DateTime.Now returns a DateTime object with the Millisecond property set to zero. Also the Ticks property is rounded accordingly. How to get better time accuracy? Remember, that my code runs on on the Compact Framework, version 3.5. I use a HTC touch Pro 2 device. Based on the answer from MusiGenesis i have created the following class which solved this problem: /// <summary> /// A more precisely implementation of some DateTime properties on mobile devices. /// </summary> /// <devdoc>Tested on a HTC Touch Pro2.</devdoc> public static class DateTimePrecisely { /// <summary> /// Remembers the start time when this model was created. /// </summary> private static DateTime _start = DateTime.Now; /// <summary> /// Remembers the system uptime ticks when this model was created. This /// serves as a more precise time provider as DateTime.Now can do. /// </summary> private static int _startTick = Environment.TickCount; /// <summary> /// Gets a DateTime object that is set exactly to the current date and time on this computer, expressed as the local time. /// </summary> /// <returns></returns> public static DateTime Now { get { return _start.AddMilliseconds(Environment.TickCount - _startTick); } } }

    Read the article

  • Geolocation Firefox accurate than iPhone Safari?

    - by johnz
    I just tested Geolocation on Firefox 3.6 and iPhone Safari (os 3.1.3), the result is interesting, firefox is more accurate than safari. any one got idea how to make iPhone Safari result more accurate. this is the code for testing: navigator.geolocation.getCurrentPosition(handler, {enableHighAccuracy: true}); function handler(location) { var message = document.getElementById("message"); message.innerHTML = "<img src='http://maps.google.com/staticmap?sensor=true&center=" + location.coords.latitude + "," + location.coords.longitude + "&size=300x300&maptype=street&zoom=16&key=ABQIAAAAZrVtlT2df2pkfI_RZB_6WBRWTAkRKJS7h1XjKaOTqACHuw1n0BT5cATkkKFnZNGHmrwUw9IilQK0Eg' />"; message.innerHTML+="<p>Longitude: " + location.coords.longitude + "</p>"; message.innerHTML+="<p>Latitude: " + location.coords.latitude + "</p>"; message.innerHTML += "<p>Accuracy: " + location.coords.accuracy + "</p>"; // call the function with my current lat/lon getPlaceFromFlickr(location.coords.latitude, location.coords.longitude, 'output'); } . . test from here

    Read the article

  • jQuery.width() & DOM refresh

    - by o_O Tync
    My script dynamically creates a <ul> width left-floating <li>s inside: it's a paginator. Afterwards, the script measures width of all <li>s and summs them up. The problem is that after the nodes are injected into the document — the browser refreshed DOM and applies CSS styles which takes a while. It has a negative effect on my script: when these operations are not complete before I measure the width — my script gets a wrong value. If I perform the measure in a second — everything is ok. The thing I'm looking for is a way to detect the moment when the <ul> is fully drawn, styles applied and the width has stabilizes. Or at least a way to detect every dimensions changes. If there's a way to detect width stabilization — I would do the measuring right after it to get the correct values. P.S. Why I need this. My paginator's left-floating <li> items tend to move to the next line when the <ul> tries to become wider than the page itself. Even though most of <li>s are invisible because of parent <div>'s width restriction: div { width: 500px; overflow: hidden; } div ul { width: 100%; white-space: nowrap; } div ul li { display: block; float: left; } they still go down unless I specify the actual summed width of the <ul> with the script.

    Read the article

  • SVM Classification - minimum number of input sets for each class

    - by Amol Joshi
    Im trying to build an app to detect images which are advertisements from the webpages. Once I detect those Ill not be allowing those to be displayed on the client side. From the help that I got here in stackoverflow, I thought SVM is the best approach to my aim. So, I have coded SVM and an SMO myself. The dataset which I have got from UCI data repository has 3280 instances ( Link to Dataset- http://archive.ics.uci.edu/ml/datasets/Internet+Advertisements )where around 400 of them are from class representing Advertisement images and rest of them representing non-advertisement images. Right now Im taking the first 2800 input sets and training the SVM. But after looking at the accuracy rate I realised that most of those 2800 input sets are from non-advertisement image class. So Im getting very good accuracy for that class. So what can I do here? About how many input set shall I give to SVM to train and how many of them for each class? Thanks. Cheers. ( Basically made a new question because the context was different from my previous question. http://stackoverflow.com/questions/1991113/optimization-of-neural-network-input-data )

    Read the article

  • Accurate least-squares fit algorithm needed

    - by ggkmath
    I've experimented with the two ways of implementing a least-squares fit (LSF) algorithm shown here. The first code is simply the textbook approach, as described by Wolfram's page on LSF. The second code re-arranges the equation to minimize machine errors. Both codes produce similar results for my data. I compared these results with Matlab's p=polyfit(x,y,1) function, using correlation coefficients to measure the "goodness" of fit and compare each of the 3 routines. I observed that while all 3 methods produced good results, at least for my data, Matlab's routine had the best fit (the other 2 routines had similar results to each other). Matlab's p=polyfit(x,y,1) function uses a Vandermonde matrix, V (n x 2 matrix) and QR factorization to solve the least-squares problem. In Matlab code, it looks like: V = [x1,1; x2,1; x3,1; ... xn,1] % this line is pseudo-code [Q,R] = qr(V,0); p = R\(Q'*y); % performs same as p = V\y I'm not a mathematician, so I don't understand why it would be more accurate. Although the difference is slight, in my case I need to obtain the slope from the LSF and multiply it by a large number, so any improvement in accuracy shows up in my results. For reasons I can't get into, I cannot use Matlab's routine in my work. So, I'm wondering if anyone has a more accurate equation-based approach recommendation I could use that is an improvement over the above two approaches, in terms of rounding errors/machine accuracy/etc. Any comments appreciated! thanks in advance.

    Read the article

  • Calculating Nearest Match to Mean/Stddev Pair With LibSVM

    - by Chris S
    I'm new to SVMs, and I'm trying to use the Python interface to libsvm to classify a sample containing a mean and stddev. However, I'm getting nonsensical results. Is this task inappropriate for SVMs or is there an error in my use of libsvm? Below is the simple Python script I'm using to test: #!/usr/bin/env python # Simple classifier test. # Adapted from the svm_test.py file included in the standard libsvm distribution. from collections import defaultdict from svm import * # Define our sparse data formatted training and testing sets. labels = [1,2,3,4] train = [ # key: 0=mean, 1=stddev {0:2.5,1:3.5}, {0:5,1:1.2}, {0:7,1:3.3}, {0:10.3,1:0.3}, ] problem = svm_problem(labels, train) test = [ ({0:3, 1:3.11},1), ({0:7.3,1:3.1},3), ({0:7,1:3.3},3), ({0:9.8,1:0.5},4), ] # Test classifiers. kernels = [LINEAR, POLY, RBF] kname = ['linear','polynomial','rbf'] correct = defaultdict(int) for kn,kt in zip(kname,kernels): print kt param = svm_parameter(kernel_type = kt, C=10, probability = 1) model = svm_model(problem, param) for test_sample,correct_label in test: pred_label, pred_probability = model.predict_probability(test_sample) correct[kn] += pred_label == correct_label # Show results. print '-'*80 print 'Accuracy:' for kn,correct_count in correct.iteritems(): print '\t',kn, '%.6f (%i of %i)' % (correct_count/float(len(test)), correct_count, len(test)) The domain seems fairly simple. I'd expect that if it's trained to know a mean of 2.5 means label 1, then when it sees a mean of 2.4, it should return label 1 as the most likely classification. However, each kernel has an accuracy of 0%. Why is this? On a side note, is there a way to hide all the verbose training output dumped by libsvm in the terminal? I've searched libsvm's docs and code, but I can't find any way to turn this off.

    Read the article

  • CIL and JVM Little endian to big endian in c# and java

    - by Haythem
    Hello, I am using on the client C# where I am converting double values to byte array. I am using java on the server and I am using writeDouble and readDouble to convert double values to byte arrays. The problem is the double values from java at the end are not the double values at the begin giving to c# writeDouble in Java Converts the double argument to a long using the doubleToLongBits method , and then writes that long value to the underlying output stream as an 8-byte quantity, high byte first. DoubleToLongBits Returns a representation of the specified floating-point value according to the IEEE 754 floating-point "double format" bit layout. The Program on the server is waiting of 64-102-112-0-0-0-0-0 from C# to convert it to 1700.0 but he si becoming 0000014415464 from c# after c# converted 1700.0 this is my code in c#: class User { double workingStatus; public void persist() { byte[] dataByte; using (MemoryStream ms = new MemoryStream()) { using (BinaryWriter bw = new BinaryWriter(ms)) { bw.Write(workingStatus); bw.Flush(); bw.Close(); } dataByte = ms.ToArray(); for (int j = 0; j < dataByte.Length; j++) { Console.Write(dataByte[j]); } } public double WorkingStatus { get { return workingStatus; } set { workingStatus = value; } } } class Test { static void Main() { User user = new User(); user.WorkingStatus = 1700.0; user.persist(); } thank you for the help.

    Read the article

  • More than one location provider at same time

    - by Rabarama
    I have some problems with location systems. I have a service that implements locationlistener. I want to get the best location using network when possible, gps if network is not enough accurate (accuracy greater than 300mt). The problem is this. I need location (accurate if possible, inaccuarte otherways) every 5 minutes. I start with a : LocationManager lm=(LocationManager)getApplicationContext().getSystemService(LOCATION_SERVICE); Criteria criteria = new Criteria(); criteria.setAccuracy(Criteria.ACCURACY_COARSE); criteria.setAltitudeRequired(false); criteria.setBearingRequired(false); String provider=lm.getBestProvider(criteria, true); if(provider!=null){ lm.requestLocationUpdates( provider,5*60*1000,0,this); In "onLocationChanged" i listen to locations and when i get a location with accuracy greater than 300mt, i want to change to gps location system. If I remove allupdates and then request for gps updates, like this: lm.removeUpdates((android.location.LocationListener) this); Criteria criteria = new Criteria(); criteria.setAccuracy(Criteria.ACCURACY_FINE); criteria.setAltitudeRequired(false); criteria.setBearingRequired(false); String provider=lm.getBestProvider(criteria, true); if(provider!=null){ lm.requestLocationUpdates( provider,5*60*1000,0,this); } system stops waiting for gpsupdate, and if i'm in a close room it can stay without location updates for hours, ignoring timeupdate indications. Is there a way to tell locationprovider to switch to network if gps is not giving a location in "x" seconds? or how to understand when gps is not localizing? or if i requestlocationupdates from 2 providers at same time (network and gps), can be a problem? Any suggestion?

    Read the article

  • Matlab-Bisection-Newton-Secant , finding roots?

    - by i z
    Hello and thanks in advance for your possible help ! Here's my problem: I have 2 functions f1(x)=14.*x*exp(x-2)-12.*exp(x-2)-7.*x.^3+20.*x.^2-26.*x+12 f2(x)=54.*x.^6+45.*x.^5-102.*x.^4-69.*x.^3+35.*x.^2+16.*x-4 Make the graph for those 2, the first one in [0,3] and the 2nd one in [-2,2]. Find the 3 roots with accuracy of 6 decimal digits using a) bisection ,b) newton,c)secant.For each root find the number of iterations that have been made. For Newton-Raphson, find which roots have quadratic congruence and which don't. What is the main common thing that roots with no quadratic congruence (Newton's method)? Why ? Excuse me if i ask silly things, but i'm asked to do this with no Matlab courses and I'm trying to learn it myself. There are many issues i have with this exercise . Questions : 1.I only see 2 roots in the graph for the f1 function and 4-5 (?) roots for the function f2 and not 3 roots as the exercise says. Here's the 2 graphs : http://postimage.org/image/cltihi9kh/ http://postimage.org/image/gsn4sg97f/ Am i wrong ? Do both have only 3 roots in [0,3] and [-2,2] ? Concerning the Newton's method , how am i supposed to check out which roots have quadratic congruence and which not??? Accuracy means tolerance e=10^(-6), right ?

    Read the article

  • compact Number formatting behavior in Java (automatically switch between decimal and scientific notation)

    - by kostmo
    I am looking for a way to format a floating point number dynamically in either standard decimal format or scientific notation, depending on the value of the number. For moderate magnitudes, the number should be formatted as a decimal with trailing zeros suppressed. If the floating point number is equal to an integral value, the decimal point should also be suppressed. For extreme magnitudes (very small or very large), the number should be expressed in scientific notation. Alternately stated, if the number of characters in the expression as standard decimal notation exceeds a certain threshold, switch to scientific notation. I should have control over the maximum number of digits of precision, but I don't want trailing zeros appended to express the minimum precision; all trailing zeros should be suppressed. Basically, it should optimize for compactness and readability. 2.80000 - 2.8 765.000000 - 765 0.0073943162953 - 0.00739432 (limit digits of precision—to 6 in this case) 0.0000073943162953 - 7.39432E-6 (switch to scientific notation if the magnitude is small enough—less than 1E-5 in this case) 7394316295300000 - 7.39432E+6 (switch to scientific notation if the magnitude is large enough—for example, when greater than 1E+10) 0.0000073900000000 - 7.39E-6 (strip trailing zeros from significand in scientific notation) 0.000007299998344 - 7.3E-6 (rounding from the 6-digit precision limit causes this number to have trailing zeros which are stripped) Here's what I've found so far: The .toString() method of the Number class does most of what I want, except it doesn't upconvert to integer representation when possible, and it will not express large integral magnitudes in scientific notation. Also, I'm not sure how to adjust the precision. The "%G" format string to the String.format(...) function allows me to express numbers in scientific notation with adjustable precision, but does not strip trailing zeros. I'm wondering if there's already some library function out there that meets these criteria. I guess the only stumbling block for writing this myself is having to strip the trailing zeros from the significand in scientific notation produced by %G.

    Read the article

  • how to 'scale' these three tables?

    - by iddqd
    I have the following Tables: Players id playerName Weapons id type otherData Weapons2Player id playersID_reference weaponsID_reference That was nice and simple. Now I need to SELECT items from the Weapons table, according to some of their characteristics that i previously just packed into the otherData column (since it was only needed on the client side). The problem is, that the types have varying characteristics - but also a lot of similar data. So I'm trying to decide on the following possibilities, all of which have their pros and cons. Solution A Kill the Weapons table, and create a new table for each Weapon-Type: Weapons_Swords id bladeType damage otherData Weapons_Guns id accuracy damage ammoType otherData But how will i Link these to the Players ? create Weapons_Swords2Players, Weapons_Guns2Players for each weapon-type? (Will result in a lot more JOINS when loading the player with all his weapons...and it's also more complicated to insert a new player) or add another column to Weapons2Players called WeaponsTypeTable, then do sub-selects to the correct Weapons sub-table (seems easier, but not really right, slightly easier insert i guess) Solution B Keep the Weapons table, and add all the fields i need to it. The Problem is that then there will be NULL fields, since not all Weapon-Types use all fields (can't be right) Weapons id type accuracy damage ammoType bladeType otherData This seems to be pretty basic stuff, but i just can't decide what's best. Or is there a correct Solution C? many thanks.

    Read the article

  • New TabItem Content ActualHeight crashes Xaml Window

    - by Jack Navarro
    I am able to create new TabItems with Content dynamically to a new window by streaming the Xaml with XamlReader: NewWindow newWindow = new NewWindow(); newWindow.Show(); TabControl myTabCntrol = newWindow.FindName("GBtabControl") as TabControl; StringReader stringReader = new StringReader(XamlGrid); XmlReader xmlReader = XmlReader.Create(stringReader); TabItem myTabItem = new TabItem(); myTabItem.Header = qDealName; myTabItem.Content = (UIElement)XamlReader.Load(xmlReader); myTabCntrol.Items.Add(myTabItem); This works fine. It displays a new grid wrapped in a scrollviewer. The problem is access the TabItem content from the newWindow. TabItem ti = GBtabControl.SelectedItem as TabItem; string scrollvwnm = "scrollViewer" + ti.Header.ToString(); MessageBox.Show(ti.ActualHeight.ToString()); // returns 21.5 ScrollViewer scrlvwr = this.FindName(scrollvwnm) as ScrollViewer; MessageBox.Show(scrollvwnm); // Displays name double checked for accuracy MessageBox.Show(scrlvwr.ActualHeight.ToString()); //Crashes ScrollViewer scrlvwr = ti.FindName(scrollvwnm) as ScrollViewer; MessageBox.Show(scrollvwnm); // Displays name double checked for accuracy MessageBox.Show(scrlvwr.ActualHeight.ToString()); //Also Crashes Is there a method to refresh UI in XAML so the new window is able to access the newly loaded tab item content? Thanks

    Read the article

  • A couple of questions about NHibernate's GuidCombGenerator

    - by Eyvind
    The following code can be found in the NHibernate.Id.GuidCombGenerator class. The algorithm creates sequential (comb) guids based on combining a "random" guid with a DateTime. I have a couple of questions related to the lines that I have marked with *1) and *2) below: private Guid GenerateComb() { byte[] guidArray = Guid.NewGuid().ToByteArray(); // *1) DateTime baseDate = new DateTime(1900, 1, 1); DateTime now = DateTime.Now; // Get the days and milliseconds which will be used to build the byte string TimeSpan days = new TimeSpan(now.Ticks - baseDate.Ticks); TimeSpan msecs = now.TimeOfDay; // *2) // Convert to a byte array // Note that SQL Server is accurate to 1/300th of a millisecond so we divide by 3.333333 byte[] daysArray = BitConverter.GetBytes(days.Days); byte[] msecsArray = BitConverter.GetBytes((long) (msecs.TotalMilliseconds / 3.333333)); // Reverse the bytes to match SQL Servers ordering Array.Reverse(daysArray); Array.Reverse(msecsArray); // Copy the bytes into the guid Array.Copy(daysArray, daysArray.Length - 2, guidArray, guidArray.Length - 6, 2); Array.Copy(msecsArray, msecsArray.Length - 4, guidArray, guidArray.Length - 4, 4); return new Guid(guidArray); } First of all, for *1), wouldn't it be better to have a more recent date as the baseDate, e.g. 2000-01-01, so as to make room for more values in the future? Regarding *2), why would we care about the accuracy for DateTimes in SQL Server, when we only are interested in the bytes of the datetime anyway, and never intend to store the value in an SQL Server datetime field? Wouldn't it be better to use all the accuracy available from DateTime.Now?

    Read the article

  • IE7 is clipping my text. How do I adjust its attitude?

    - by Emiel
    Hi All, A few days ago I re-skinned my website, http://emle.nl. Development of this skin was primarily done using safari, and as expected, it all renders fine using firefox and opera. I've had to make a few small tweaks for IE7, but nothing much, except for one problem... The date indicators for a post are cut off in IE. This problem seems to occur only on nested span tags inside a left floating div. I think I need the floating div's in order to layout text on the left and the right side of the screen. Anyhow, I've summarized it into a small test case, located at http://emle.nl/test.html. In the different browsers, it looks like this. Of course safari and firefox get this right: Do any of you know how to stop IE7 from clipping my text? Edit: I have sort of given up on this problem. My scripts now check for IE7 and feed it somewhat simplified HTML that its limited engine can handle. It works in IE8, so, for now, just the special case for IE7 will have to do...

    Read the article

  • Binary files printing and desired precision

    - by yCalleecharan
    Hi, I'm printing a variable say z1 which is a 1-D array containing floating point numbers to a text file so that I can import into Matlab or GNUPlot for plotting. I've heard that binary files (.dat) are smaller than .txt files. The definition that I currently use for printing to a .txt file is: void create_out_file(const char *file_name, const long double *z1, size_t z_size){ FILE *out; size_t i; if((out = _fsopen(file_name, "w+", _SH_DENYWR)) == NULL){ fprintf(stderr, "***> Open error on output file %s", file_name); exit(-1); } for(i = 0; i < z_size; i++) fprintf(out, "%.16Le\n", z1[i]); fclose(out); } I have three questions: Are binary files really more compact than text files?; If yes, I would like to know how to modify the above code so that I can print the values of the array z1 to a binary file. I've read that fprintf has to be replaced with fwrite. My output file say dodo.dat should contain the values of array z1 with one floating number per line. I have %.16Le up in my code but I think that %.15Le is right as I have 15 precision digits with long double. I have put a dot (.) in the width position as I believe that this allows expansion to an arbitrary field to hold the desired number. Am I right? As an example with %.16Le, I can have an output like 1.0047914240730432e-002 which gives me 16 precision digits and the width of the field has the right width to display the number correctly. Is placing a dot (.) in the width position instead of a width value a good practice? Thanks a lot...

    Read the article

< Previous Page | 18 19 20 21 22 23 24 25 26 27 28 29  | Next Page >