I'm importing text items to Google's AutoML. Each row contains around 5000 characters and I'm adding 70K of these rows. This is a multi-label data set. There is no progress bar or indication of how long this process will take. Its been running for a couple of hours. Is there any way to calculate time remaining or total estimated time. I'd like to add additional data sets, but I'm worried that this will be a very long process before the training even begins. Any sort of formula to create even a semi-wild guess would be great.
-Thanks!
I don't think that's possible today, but I filed a feature request [1] that you can follow for updates. I asked for both training and importing data, as for training it could be useful too.
I tried training with 50K records (~ 300 bytes/record) and the load took more than 20 mins after which I killed it. I retried with 1K, which ran for 20 mins and then emailed me an error message saying I had multiple labels per input (yes, so what? training data is going to have some of those) and I had >100 labels. I simplified the classification buckets and re-ran. It took another 20 mins and was successful. Then I ran 'training' which took 3 hours and billed me $11. That maps to $550 for 50K recs, assuming linear behavior. The prediction results were not bad for a first pass, but I got the feeling that it is throwing a super large neural net at the problem. Would help if they said what NN it was and its dimensions. They do say "beta" :)
don't wast your time trying to using google for text classification. I am a GCP hard user but microsoft LUIS is far better, precise and so much faster that I can't believe that both products are trying to solve same problem.
Luis has a much better documentation, support more languages, has a much better test interface, way faster.. I don't know if is cheaper yet because the pricing model is different but we are willing to pay more.
Related
While training my predictor I came across this error and I got stuck how to fix it.
I have two data-series, a "Target time-series data" with 9234 rows and a single "item_id" and a second one that is "Related time-series data" with the same number of rows as I only have a single id.
I'm setting de data with a window of 180 days, what is exactly the difference between the second and the first number that has appeared on the error, 9414 - 9234 = 180.
We were unable to train your predictor.
Please ensure there are no missing values for any items in the related time series, All items need data until 2020-03-15 00:00:00.0. For example, following items have missing data: item: brl only has 9234/9414 required datapoints starting 1994-06-07 00:00:00.0, please refer to documentation for additional details.
Once my data don't have missing data and it's on a daily basis why is it returning this error?
My data starts on 1994-06-07 and ends on 2019-09-17. Why should I have 9414 data points rather than 9234?
Should I take out 180 days in my "Target time-series data"?
The future values of the related time-series data must be known.
Example of a good related-time series: You know past and future days in which marketing has or will send email newsletters promoting the product you're forecasting. You can use this data as a related-time series.
Example of a bad related-time series: You notice that Google searches for your brand correlated with the sale of your product. As a result you want to use it as a related-time series. Since you don't know how many searches will occur in the future, so you can't use this as a related time series.
In you case, You have TARGET_TIME_SERIES data for 9414 days and you want to predict demand for the next 180 days. That means your RELATED_TIME_SERIES data should be 9594 days.
Edit: I have not tested this with amazon's forecasting product. I'm basing my answer on working with Facebook Prophet (which is one of the models amazon forcast uses). Please let me know if my solution worked.
I have a dataset from which I would like to detect recurring patterns (i.e: daily, weekly, monthly). The dataset only contains a time stamp (datetime), and the spacing is non-uniform.
The observations in the data reflect the exact time when this one person passes my window. He does this several times a day (on a single day he walks by my window approx 10-30 times), and I am trying to see, if there is any pattern (there might also be some seasonality, sudden changes in previous behavior and other interesting stuff going on).
Does anyone have a suggestion for a statistical model/approach that might be helpful in figuring out if there is any pattern in this behavior? Hopefully, I’ll be able to predict when he will pass my window again ;)
How would you approach this?
Any help would really be appreciated.
I'm using the AWS Machine Learning regression to predict the waiting time in a line of a restaurant, in a specific weekday/time.
Today I have around 800k data.
Example Data:
restaurantID (rowID)weekDay (categorical)time (categorical)tablePeople (numeric)waitingTime (numeric - target)1 sun 21:29 2 23
2 fri 20:13 4 43
...
I have two questions:
1)
Should I use time as Categorical or Numeric?
It's better to split into two fields: minutes and seconds?
2)
I would like in the same model to get the predictions for all my restaurants.
Example:
I expected to send the rowID identifier and it returns different predictions, based on each restaurant data (ignoring others data).
I tried, but it's returning the same prediction for any rowID. Why?
Should I have a model for each restaurant?
There are several problems with the way you set-up your model
1) Time in the form you have it should never be categorical. Your model treats times 12:29 and 12:30 as two completely independent attributes. So it will never use facts it learn about 12:29 to predict what's going to happen at 12:30. In your case you either should set time to be numeric. Not sure if amazon ML can convert it for you automatically. If not just multiply hour by 60 and add minutes to it. Another interesting thing to do is to bucketize your time, by selecting which half hour or wider interval. You do it by dividing (h*60+m) by some number depending how many buckets you want. So to try 120 to get 2 hr intervals. Generally the more data you have the smaller intervals you can have. The key is to have a lot of samples in each bucket.
2) You should really think about removing restaurantID from your input data. Having it there will cause the model to over-fit on it. So it will not be able to make predictions about restaurant with id:5 based on the facts it learn from restaurants with id:3 or id:9. Having restaurant id there might be okay if you have a lot of data about each restaurant and you don't care about extrapolating your predictions to the restaurants that are not in the training set.
3) You never send restaurantID to predict data about it. The way it usually works you need to pick what are you trying to predict. In your case probably 'waitingTime' is most useful attribute. So you need to send weekDay, time and number of people and the model will output waiting time.
You should think what is relevant for the prediction to be accurate, and you should use your domain expertise to define the features/attributes you need to have in your data.
For example, time of the day, is not just a number. From my limited understanding in restaurant, I would drop the minutes, and only focus on the hours.
I would certainly create a model for each restaurant, as the popularity of the restaurant or the type of food it is serving is having an impact on the wait time. With Amazon ML it is easy to create many models as you can build the model using the SDK, and even schedule retraining of the models using AWS Lambda (that mean automatically).
I'm not sure what the feature called tablePeople means, but a general recommendation is to have as many as possible relevant features, to get better prediction. For example, month or season is probably important as well.
In contrast with some answers to this post, I think resturantID helps and it actually gives valuable information. If you have a significant amount of data per each restaurant then you can train a model per each restaurant and get a good accuracy, but if you don't have enough data then resturantID is very informative.
1) Just imagine what if you had only two columns in your dataset: restaurantID and waitingTime. Then wouldn't you think the restaurantID from the testing data helps you to find a rough waiting time? In the simplest implementation, your waiting time per each restaurantID would be the average of waitingTime. So definitely restaurantID is a valuable information. Now that you have more features in your dataset, you need to check if restaurantID is as effective as the other features or not.
2) If you decide to keep restaurantID then you must use it as a categorical string. It should be a non-parametric feature in your dataset and maybe that's why you did not get a proper result.
On the issue with day and time I agree with other answers and considering that you are building your model for the restaurant, hourly time may give a more accurate result.
I have to partition the data by a date field in it. I am doing it using Partition Transform.
When I divide yearly data by month, Partition returns a Pcollectionlist which has 12 pcollection. This works fine.
When I have to divide it by day. I will have to create 1*12*31 Pcollection in PcollectionList. This throughs Heap space error. I tried only for 2 months data. That is,
a PcollectionList of 2*31 Pcollection
I tried using n1-highmem-4 and n1-highmem-8 machines with more than 10 workers. Still it throughs Heap space error. I am testing with only 2.0 MiB file. So I believe data size should not be a problem. The screen shots are below.
Please help me to fix this. Or a work around to my solution is also most welcome.
Thanks in advance.
It sounds like you're trying to get time-based divisions of your data. Have you looked at windowing? It should allow you to do monthly/daily/hourly windowing without needing to perform the partition. If windowing isn't applicable, could you explain why you need to partition by day?
How are you consuming the partitioned results? You may be running into a known bug with pipelines with many sinks running into OOM errors due to the byte buffers for each of the sinks.
I am currently working on a project that requires load testing of web services.
One of the services is being called 60,000 times in the production during Busy-Day/Busy-HR.
{PerfTest Env=PROD}
Input Account Number
Output AccountDetails
Do I really need 60,000 unique account numbers(TEST DATA) for this loadrunner script to simulate the production scenario?
If unique data is required, for endurance test I will have to prepare lot of test data for each web service.
If I don't get that much test data, what is the chance of Load Test being affected due to Application Server Cache mechanism??
Can somebody help me?
Thanks
Ram
Are you simulating a day or the highest volume hour in the last year? This can help you to shape the amount of data that you need. Rarely would you start with a 24 hour test. Instead you would be looking at your high water test of an hour with a ramp up and ramp down, so you would need approximately 1.333* your high water hour's worth of data.
So this can drop your 60K to (potentially) 20K(?) I am making an assumption that your worst hour over the last year is somewhere around 1/3 of your traditional day. I have observed this pattern over and over again in different environments over the past two decades. You will want to objectively verify this with log data or query data to support the number in your environment.
Next up, how many of these inquiries are actually unique? You are really going to need a log of the queries across a day (or your high water hour) to determine this. Log processing tools such as Microsoft Logparser or Splunk/Splunk Storm can help you to pull the observed distribution of unique account references within your data, including counts of those which are multiple. Once you know this you can simply use a data file with a fixed block size for each user for unique data and once the data is exhausted the user exits.