Stack Exchange Network

Stack Exchange network consists of 174 Q&A communities including Stack Overflow, the largest, most trusted online community for developers to learn, share their knowledge, and build their careers.

Visit Stack Exchange

Questions tagged [machine-learning]

Methods and principles of building "computer systems that automatically improve with experience."

5
votes
0answers
31 views

How to deal with possible data leakage in time series data?

I have historical consumer data who have taken out a loan at some point in time. The task is to predict if a consumer will default when requesting a loan. My issue is that for some customer in the ...
5
votes
0answers
13 views

How to methodologically show that a given 'time-series/sequential' data is not really sequential?

I have an apparent time-series/sequencial (supervised: multi-class classification problem) dataset with each data-point time-stamped. However, some domain intuition tells me that the data is static ...
5
votes
0answers
219 views

Clustering for high dimensional data

I am having a data set with 52 variables .Most of them are having zeros it resembles like sparse matrix .How to cluster this kind of data and is there any special types of clustering? Here I am ...
5
votes
0answers
495 views

document classification : what is the difference betwen fasttext and DANs ?

I came across two interesting papers which describe promising approaches for document classification using word embedding. [1] the first one is the fasttext algorithm described in the paper : Bag of ...
5
votes
0answers
397 views

I am struggling with Time Series pattern recognition and classification problem?

I have some sensor data and it is labelled. Now I want to know how can I extract features from time series using DFT, DWT and HAAR transforms. I know that transformations above transform a signal ...
5
votes
0answers
107 views

How to model segmentation of a sequence to similar parts?

I guess LSTM is good for sequence modeling but how would you model "clustering" with it? Meaning, the input is a sequence and the output is labels with similar properties (I have labeled data). For ...
5
votes
0answers
81 views

Is Minimax Linkage a Lance-Williams hierarchical clustering?

I found the following article on "Hierarchical Clustering With Prototypes via Minimax Linkage". It is stated in Property 6 that Minimax linkage cannot be written using Lance–Williams updates. A ...
4
votes
0answers
479 views

Input for LSTM for financial time series directional prediction

I'm working on using an LSTM to predict the direction of the market for the next day. My question concerns the input for the LSTM. My data is a financial time series $x_1 \ldots x_t$ where each $x_i$...
4
votes
0answers
439 views

AUC computation on multilabel classification

I'm using Tensorflow for an auto-tagging task on audio clips. The problem is actually a multilabel classification problem meaning that each clip can have multiple tags at the same time. Regarding ...
4
votes
0answers
723 views

AdaBoost implementation and tuning for high dimensional feature space in R

I am trying to implement the AdaBoost.M1 algorithm (trees as base-learners) to a data set with a large feature space (~ 20.000 features) and ~ 100 samples in R. There exists a variety of different ...
4
votes
0answers
217 views

how to propagate error from convolutional layer to previous layer?

I've been trying to implement a simple convolutional neural network. But I've been stuck at this problem for over a week. To be specific, assume there are 3 layers in a convolutional pass, marked as ...
4
votes
0answers
174 views

Human Activity Recognition - How to process gyroscope and magnetometer data

I've found many articles (like this one) on HAR that mention using the gyroscope and magnetometer because it improves the precision of HAR performed just with the accelerometer. However no one ...
4
votes
0answers
113 views

Heterodox use of Deep Learning to find hidden patterns

I would appreciate your comments/help about a strategy I am applying in one of my analysis. In short, my case is: 1) My data have biological origin, collected in a period of 120s, from a subject ...
3
votes
0answers
17 views

Scaling does not speed up the SVM model

I tried to standardize the training data with samples of 629,145 rows and 24 features: ...
3
votes
0answers
29 views

Why I am getting prediction score 1 i.e. 100%

I am reading few parameters and trying to predict target value using Linear regression and GB. Surpicingly I am getting score = 1 on test data. How come? Can anyone tell me whats wrong with this code? ...
3
votes
0answers
24 views

How to deal with count data in random forest

I am working on a classification model where my target class is a biased class with the class shape as 0 1 20694 101 Most of my features are the ...
3
votes
0answers
32 views

Training value neural network AlphaGo style

I have been trying to replicate the results obtained by AlphaGo following their supervise learning protocol. The papers specify that they use a network that has two heads: a value head that predicts ...
3
votes
0answers
45 views

RNN-based Predictions of Sine Waves with Frequency Different From Training Data

I am wondering if I can generate a sine wave with a frequency different from training data using RNN. For example, Using two training data of two time series, say 0[sec] ~ 10[sec] each: sin(t) and ...
3
votes
0answers
38 views

Robustness of ML Model in question

While trying to emulate a ML model similar to the one described in this paper, I seemed to eventually get good clustering results on some sample data after a bit of tweaking. By "good" results, I mean ...
3
votes
0answers
29 views

Deep Learning: Does starting the training on a smaller subset of the data make sense?

I trained a deep neural network with a small subset of my data, which allowed me to go through many epochs in a short amount of time and allowed the model to perform reasonably, then I gave it the ...
3
votes
0answers
51 views

How to best estimate the coefficients of a confusion matrix in case of strong class imbalance?

I have a trained binary classifier (forget about how this was trained and think of it as a magical black box) and I would like to measure its classification performance (e.g. compute a confusion ...
3
votes
0answers
18 views

Tuning Lexicon Sentiment-values Using Machine-learning

I'm constructing a sentiment-analysis model using the lexicon-based approach, and wondering if I can tune the weights of each word-feature in the lexicon using machine-learning. Is this achievable ...
3
votes
0answers
87 views

Machine Learning & Image Recognition: How to start?

I've been a full stack web developer for 15 years now and would like to be involved in machine learning. There is already a specific scenario for this: We have a database with several million products ...
3
votes
0answers
43 views

Predicting change of shapes/coordinates

I'm trying to find a way to predict/calculate how a shape (e.g. outline of a glacier) will change in the future—based on its history (previous shape) and additional factors (e.g. Δtemperature). In my ...
3
votes
0answers
176 views

How to properly represent a tic tac toe board to a CNN?

I'm figuring out how to manipulate convolutional neural networks (CNN) in python and I want to apply this kind of NN to an agent player that plays tic tac toe. I know that's weird and the problem ...
3
votes
0answers
33 views

Hochreiter LSTM (p. 4): Maximal values of logistic sigmoid derivative times weight

My questions follow the below page 4 excerpt from Hochreiter's LSTM paper: If $f_{l_{m}}$ is the logistic sigmoid function, then the maximal value of $f^\prime_{l_{m}}$ is 0.25. If $y^{l_{m-1}}$ ...
3
votes
0answers
902 views

ValueError: operands could not be broadcast together with shapes (60002,39) (38,) during pca.transform

I am trying to solve the San Francisco Crime Problem on Kaggle. To begin with, here is my code: ...
3
votes
0answers
25 views

Formula to calculate size of Capsule output similar to the formula for CNN?

Is there any formula to find the output dimensions of a capsule network similar to that of a Convolutional Neural Network? For Example: In CNN, we know that ...
3
votes
0answers
87 views

Confidence value in AdaBoost?

I read this introduction about AdaBoost (http://www.cs.man.ac.uk/~nikolaon/~nikolaon_files/Introduction_to_AdaBoost.pdf), and am curious why confidence for each model is defined as $$\alpha_j=\frac{...
3
votes
0answers
42 views

Deriving diagonal approximation of Hessian in a neural network

Consider the equations relating to the diagonal approximation for the hessian matrix for a neural network in "Pattern Recognition and Machine Learning - Christopher Bishop" (on pg. 250 eq. 5.80) $\...
3
votes
0answers
101 views

How to do give input to CNN when doing a text processing?

As a signal processing engineering and being new to NLP, I am confused with giving input to CNN network. With my knowledge of CNN, I am trying to build a classifier for ethnicity with inputs as text ...
3
votes
0answers
53 views

make prediction with a time serie

I want to build a model that can detect which driver is driving now the car based on a dataset that contains 20 driver records for 3600s each driver ( the dataset contains all the car sensors values ...
3
votes
0answers
590 views

Help with the following error: Variable already exists, disallowed. Did you mean to set reuse=True in VarScope?

I am not sure how to handle this error. This is from an RNN tutorial found here. I vaguely understand that the variables need to be able to be reused, but I don't know how to implement this fix. ...
3
votes
0answers
2k views

Why we should not feed LDA with tfidf

Can someone explain why we can not feed LDA topic model with TFIDF? What is wrong with this approach conceptually?
3
votes
0answers
406 views

Sales Dataset to determine best model for predicting future sales

We have a set of products in which we are trying to determine which products we should continue to sell, and which products to remove from our inventory. The file contains BOTH historical sales data ...
3
votes
0answers
329 views

What causes the error in a RNN to increase late in training?

I'm training a 2-layer, 1024 node, dropout of 0.5 RNN over natural text. Specifically, I'm using karpathy's char-rnn which I found to work quite well for most of my use cases. Sometimes however, late ...
3
votes
0answers
371 views

Illustrating the dimensionality reduction done by a classification or regression model

Tl;DR: You can predict something, but how do you explain the prediction? Your usual classification/regression setup Lets say the data is a classic regression/classification problem: several ...
3
votes
0answers
71 views

Train a classifier for a game with feedback on chosen move instead of true labels

I'm having some trouble describing in one line what I want, which is probably why I haven't had much luck with Google. Say I have a game like 2048 where the possible actions each step are fixed (and ...
3
votes
0answers
82 views

Is there an interface for bootstrap sampling in rattle in R?

I am using rattle in R for predictive models and am trying to see whether there is a difference in different sampling methods. The split function at the start of rattle for splitting into training and ...
3
votes
0answers
68 views

Various algorithms performance in a problem and what can be deduced about data and problem?

HI I am currently trying to apply various algorithms to a classification problem to assess which could be better and then try to fine tune the bests of the first approach. I am a beginner so I use ...
2
votes
0answers
9 views

Should one normalize the frequency values when feeding it as an input to machine learning model?

Consider an unsupervised data. The data is in the form of a csv file( I am using pandas dataframe for this). Its a web page data at different time steps and the way I am converting data to be fed to ...
2
votes
0answers
10 views

Sequence to sequence LSTM is “slow” to observe changes. Can this be improved?

I have some multivariate time series, and I'm trying to teach an LSTM with simulated data what is good and what is bad, so the input is a timeseries with a label for every timestep, i.e. [1,1,1,1,1,1,...
2
votes
0answers
10 views

How to implement Moving window with LSTM for Time Series Prediction?

I am trying to implement a moving window in my dataset. The window size=14 (for instance).After implemntinf sliding window how to prepare inputs and outputs for netwok? ...
2
votes
0answers
41 views

How to handle preprocessing (StandardScaler, LabelEncoder) when using data generator to train?

So, I have a dataset that is too big to load into memory all at once. Therefore I want to use a generator to load batches of data to train on. In this scenario, how do I go about performing scaling ...
2
votes
0answers
29 views

Regularization in Embedding models?

What is the best way to regularize latent embeddings, I have two solution in my mind but I'm not sure which one to use over other. In batch-wise training regularize over the whole embedding matrix, ...
2
votes
0answers
18 views

How to discretize numerice values with predfined ranges in weka?

I'v imported csv file into weka. one of the features have a value with minimum 0 and maximum 160. now, i want to discretize value into three range as you can see below: less than 6 > L more than 6 ...
2
votes
0answers
53 views

Regularization term in Matrix Factorization

I'm trying to build a naive recommender system using latent factor model for MovieLens dataset. From the observed set of ratings I'm trying to build a model which will decompose the sparse matrix to N ...
2
votes
0answers
57 views

Cannot interpret feed_dict key as Tensor: Tensor Tensor(“Placeholder:0”, shape=(3, 3, 3, 32), dtype=float32) is not an element of this graph

I have a working Keras model that makes predictions great in the repl but fails to load in a Flask app on accessing multiple times.. K.clear_session() and graph = tf.get_default_graph() did not work ...
2
votes
0answers
29 views

How to improve a model with a high cross validation score yet with low accuracy on unseen data?

I have a model that is based on an experiment collected on 100 subjects. We are testing the model as follows: Record raw data from the subjects For each subject, compute the feature from the raw data ...
2
votes
0answers
40 views

Target transformation for tree models

Can anybody explain why/if target variable transformations could help when dealing with tree based models? I've seen this excellent reply which explains quite well why it shouldn't affect if ...