Activity-Recognition | 手机加速度传感器数据进行人体行为识别

Following this tutorial from https://machinelearningmastery.com/how-to-develop-rnn-models-for-human-activity-recognition-time-series-classification/

I’m trying to implement CNN-LSTM Network Model on my time series data.

I'm creating a 1D CNN using tensorflow.keras, following this tutorial, with some of the concepts from this tutorial. So far modeling and training seem to be working, but I can't seem to generate a prediction. Here's an example of what I'm dealing with:

Google Fit is giving me the following exceptions when I try subscribing to TYPE_STEP_COUNT_CUMULATIVE (cumulative steps) and TYPE_STEP_COUNT_DELTA using the RecordingClient (https://developers.google.com/android/reference/com/google/android/gms/fitness/RecordingClient)

I'm struggling to figure out how to properly save a tensorflow model that was saved as a saved_model method and convert it to tensorflow.js. when I currently convert it it ends up with a weights_manifest.json containing only one set of weights.

The model I want to convert is from this jupyter notebook .RNN for Human Activity Recognition - 2D Pose Input.

The full code for building and training the graph can be found in that jupyter notebook.

I can successfully run the code:

I am trying to do a model to recognize one human movement. My activity is to recognize a person catching. I already have a record with about 260 right movements and I annoted that file with labels "catch" and "nothing" to train my model. I have another file, this file is annoted too, just to test the accuracy of the model.

I am using a CNN model based on this site. And I modified the window size to 400 (equivalent of 4s of record) So after trainning my model I have this results (my model is saved and is already trained from a previous record, that's why it gives good results in the first epochs):

I have downloaded an android app from Github (Link). But I am unable to install this app on android (I am using Linux). Could someone tell me which direction should I go? App folder Contains following files:

1. build.gradle

2. gradle.properties

3. gradlew
4. gradlew.bat
5. settings.gradle

I am an R learner. I am working on 'Human Activity Recognition' dataset from internet. It has 563 variables, the last variable being the class variable 'Activity' which has to be predicted.

I am trying to use KNN algorithm here from CARET package of R.

I have created another dataset with 561 numeric variables excluding the last 2 - subject and activity.

I ran the PCA on that and decided that I will use the top 20 PCs.

According to Google Google APIs for Android the ActivityRecognitionApi interface is deprecated and need to use the ActivityRecognitionClient interface. But other sources saying that the ActivityRecognitionClient is depricated and need to use the ActivityRecognitionApi interface insted. I am confused :(

And many other samples on the internet is using the ActivityRecognitionApi interface.

And when we using the ActivityRecognitionClient interface the periodic update is not getting only the initial time we will get the update , But he ActivityRecognitionApi interface update is getting properly.

Which one we need to use ? , If we need to uses ActivityRecognitionClient kindly provide a sample code of working Application/Code .

I am trying to write some code in python for data segmentation, here is my code it consists of two functions. I used python 3.5