Questions tagged [sums-of-squares]

sum of squares plays an important role in statistical models based on the normal distribution, like ANOVA.

Filter by
Sorted by
Tagged with
3 votes
1 answer
54 views

Cannot understand why total SS=explained SS+unexplained SS

I cannot understand why total SS=explained SS+unexplained SS because geometrically the sum of two small squares is not equal to a big square. I wish someone could explain that to me. Thank you.
  • 31
2 votes
1 answer
72 views

Different results for type I and type II ANOVA

I am performing an ANOVA for a research question. I want to to test whether "Grade", "Nim" and "Tim" has a significant influence on "value". My question is ...
  • 23
1 vote
0 answers
54 views

Is there an interpretation for this identity from sum of squares?

So yesterday my professor wrote the following on the board: $$\frac{1}{n}\sum_{i = 1}^{n}\left( x_i - \bar{X}\right)^2 = \frac{1}{n}\sum_{i = 1}^{n} x_i^2 - \bar{X}^2$$ where $\bar{X} = \frac{1}{n}\...
  • 141
1 vote
1 answer
50 views

Expected Value for Sum of Squares in Reduced Model

$\newcommand{\ol}[1]{\overline{#1}}\newcommand{\szdp}[1]{\!\left(#1\right)} \newcommand{\szdb}[1]{\!\left[#1\right]}$I am studying Design and Analysis of Experiments, 2nd Ed., by Dean, Voss, and ...
1 vote
0 answers
29 views

cannot reproduce the example for nested anova

I'm working on the example for nested ANOVA from the Handbook of Biological Statistics. Copy-pasteable data can be found at https://www.biostathandbook.com/nestedanova.html and below shows my ...
  • 131
3 votes
0 answers
155 views

What exactly do Anova sums of squares represent in lmer models?

In a lme4::lmer model, the anova function returns an sequential analysis of variance table. As explained on page 34 of http://...
  • 251
2 votes
0 answers
72 views

Interpretation of default Type III Sums in Squares in R

Many posts (i.e. here and here) discuss how the Type III sums of squares produced by car::Anova() in R are incorrect or nonsensical under R's default model parameterization, "contr.treatment"...
  • 1,171
2 votes
1 answer
52 views

How is the matrix of corrected sum of squares and products defined?

In the article Johns, S. "On identifying the population of origin of each observation in a mixture of observations from two normal populations." Technometrics 12.3 (1970): 553-563, the ...
  • 31
1 vote
1 answer
42 views

Summation of terms in survival function giving negative probability

The following expression represents a survival function for a group size (n), where survival is a percentage on a scale of 0 to 100. $$ p(x) = 100 - \beta(x \cdot x) = 100 - \beta\sum_{i=1}^{n}x_i^2 $$...
0 votes
1 answer
77 views

Does Residual Sum of Squares always decreases with an additional independent variable? [duplicate]

^ What the question says. Trying to determine whether this statement is true or false.
user avatar
0 votes
0 answers
8 views

Is there any formal proof that in R type-1 anova() run twice on 2 models with switched terms equals type-2 (looking at the last entry)?

I observed something interesting to me. Let's assume that I have an unbalanced, additive model (no interactions) with 2 terms. When I swap the order of the two terms and run type-1 ANOVA over them, ...
  • 51
0 votes
1 answer
55 views

Synonyms of Residual Sum of Squares

As a novice, I'm finding it difficult to learn statistics, partly because there are often many different words for the same thing. When I'm reading about stats, I don't realize that the thing I'm ...
-3 votes
1 answer
2k views

"Normalized mean squared error" says WHAT?

I know that mean squared error is a public and popular metric to evaluate the efficiency of the model and architecture. Also, it is the tool to evaluate the result in such if, the ...
  • 127
0 votes
0 answers
131 views

Why type III sum of squares is not recommended to use when there is significant interaction?

I did notice that in the following example, I obtained a insignificant main effect for treatment when including interaction, although the treatment is clearly effective. Any intuitive explanation? ...
  • 547
-1 votes
2 answers
241 views

Why calculate Mean Squared Error for Regression?

Once we fit the model to the points why is it necessary to find the mean squared error? What happens if I don't calculate Mean Squared Error?
  • 1
1 vote
0 answers
30 views

Which type of sum of squares corresponds to the partition of variance formula for two-way ANOVA?

Does anyone have a good reference for the formula for type I and type III sum of squares in two-way ANOVA? I learned the following formula for partition of total sum of squares in two-way ANOVA. Is it ...
  • 547
1 vote
1 answer
3k views

The number of clusters in the K-means and the within-cluster SS

Given a collection of observatons $\{X_i\}_1^N$ and prespecify the number of clusters K. The K-means solves $$ \underset{\{C_k\}_1^K}{\arg\min} \sum_{k=1}^K \sum_{i \in C_k}|| X_i - \mu_k||^2 $$ where ...
  • 122
4 votes
1 answer
276 views

Conceptual/Math/brainteaser Question: Multiple Linear Regression

This is kind of a brainteaser and I'm struggling to solve it, any ideas on approaching this would be valued: I thought about using substitution of the different x and z into the third regression ...
6 votes
1 answer
262 views

Confused: Why is lme4 changing techniques from Wald F tests to Wald Chisquare?

I've constructed some LMEMs that use dichotomous variables and their interactions as regressors, and I've become confused by the output. When I only assess a single interaction, using the code below (...
1 vote
0 answers
41 views

Fitting data to sums of squares of sinusoids

I am given some time series data $(t, y(t))$ sampled at regular intervals $t=0,s,2s,3s,\ldots,1$ for some step size $s$, obtained from a function $$ y(t) = y(t, \vec A, \vec \delta) = \sum_{i=1}^N ...
0 votes
1 answer
212 views

Which type of sums of squares does lm-function in R use?

I ran a two-way ANCOVA in R: ancova = lm(DV ~ IV1*IV2 + CV1 + CV2 + CV3, data = Data) summary.aov(ancova) Anybody know if this uses type III sums of squares? I ...
2 votes
1 answer
64 views

How unbalanced is unbalanced for factorial ANOVA?

Colleagues and I conducted an experiment with participants randomly allocated to each condition using survey software. Unfortunately, the cell sizes ended up being more unbalanced than we anticipated: ...
  • 743
10 votes
1 answer
2k views

Is least squares means (lsmeans) statistical nonsense?

I recently came accross this quote from Brian Ripley, who seems to be well-regarded as a statistician. "Some of us feel that type III sum of squares and so-called ls-means are statistical ...
  • 3,314
4 votes
1 answer
889 views

Proof that Regression Sum of Squares and Residual Sum of Squares are independent random variables

Having consulted a number of sources, I still can't find a complete proof that Regression Sum of Squares ($SS_{regression}$) and ($SS_{residual}$) are independent random variables. I'll be doubly ...
1 vote
1 answer
64 views

Shortcut for computing RSS at different split point when building regression trees?

I'm coding regression trees from scratch in R. For a given ordered predictor variable, $X$, obviously I have to compute the RSS at each unique ordered value of $X_i$. When moving over to the next ...
  • 11
0 votes
1 answer
614 views

Why contrast in ANOVA has one degree of freedom?

I thought I understood the degree of freedom after reading Wikipedia explanation, but came across the sum of squares for contrasts $\{c_i\}$ $$SS_C = \frac{(\sum_i {c_i \bar{y_i}})^2}{\sum_i c_i^2 /...
  • 11
1 vote
0 answers
474 views

Contrasts in ANOVA - need intuitions to understand its sum of squares and its degree of freedom

I have two questions about contrasts in ANOVA for testing the hypothesis that $\sum_i c_i y_i = 0$. How is the sum of squares for contrasts $\{c_i\}$, $SS_C = \frac{(\sum_i {c_i \bar{y_i}})^2}{\sum_i ...
  • 11
1 vote
0 answers
22 views

Distribution of the squared length of a multivariate normal vector [duplicate]

Let $\mathbf{X} \sim \mathcal{N}(\mathbf{\mu}, \mathbf{\Sigma})$ be a multivariate Normal variable with $d$ dimensions. I'm interested in the marginal distribution of $\|\mathbf{X}\|^2 = \sum_{i=1}^d ...
  • 111
1 vote
1 answer
409 views

In a PCA setting, is there any relationship with the sum of squares of the scores (t1) with the eigenvalue of that principal component?

When one computes the vector of scores (t1) using the principal component (p1) the data is being projected over the direction of biggest variation. One could measure the distance between the point ...
  • 65
3 votes
1 answer
397 views

What is the lowest sum of squared errors (sse) of a dataset?

Given $X \subset R^d$ a set of data points, $z \in R^d$ a vector, $SSE(X)=\sum_{i=1}^{|X|}||X_i-z||^2$, I wonder what is the best value of $z$ so that $SSE$ is the lowest? I suspect that $z$ is the ...
  • 153
1 vote
0 answers
52 views

How are terms in SS (corrected sum of squares) independent?

What does the highlighted test mean? How are the terms not independent? Can someone elaborate in detail?
0 votes
1 answer
314 views

Source for BIC with residual sum of squares?

Wikipedia states that under certain circumstances, $\mathrm{BIC} $ can be calculated as: $$\mathrm{BIC} = n \ln(\mathrm{RSS}/n) + k \ln(n),$$ where $\mathrm{RSS}$ is the residual sum of squares. ...
  • 273
1 vote
1 answer
331 views

Kruskal–Wallis one-way analysis of variance is related to what kind of regression?

One-way anova is similar to regular linear regression because both use the F-test which involves sums of squares among other reasons. Is Kruskal–Wallis one-way analysis of variance similar to some ...
user avatar
0 votes
0 answers
77 views

Difficulty with averaging corrected sample variances of different degrees of freedom:

I have a number of measurement samples of which some have 2 measurements and some have 3. I wish to make the most accurate estimation of population variance I can, and understand that ignoring data ...
  • 67
3 votes
3 answers
2k views

Residual sum of squares in a regression

I understand that in a linear regression model, the residual sum of squares will either remain same or fall with the addition of a new variable. What if the two models were $$ I \colon y_i=\...
  • 1,367
3 votes
2 answers
2k views

Simplifying the Matrix Form of the Solution to Ridge Regression

I'm trying to understand how to obtain the solution to an objective function by solving for the parameter vector $\theta$ in ridge regression. I found an example here from Naomi which takes an example ...
1 vote
0 answers
1k views

What is the correct implementation of BIC with residual sum of squares?

BIC is most often calculated by maximizing the log likelihood function. However, it is also possible to calculate BIC with residual sums of squares. This is pretty easy to find online and not an issue ...
  • 401
2 votes
0 answers
225 views

Formula for type III sum of squares of the intercept term in linear multiple regression

assume we have the regression model: $$Y = b_0 + b_1 x_1 + \dots + b_k x_k + \varepsilon $$ I know the formulas for all type III sum of squares for the regression terms except the formula for SS of ...
  • 101
6 votes
2 answers
182 views

Coefficient of determination relationship?

$R^2 = \frac{SSREG}{SSTOT}$ or $R^2 = 1-\frac{SSRES}{SSTOT}$ If $X$ is the predictor random variable for science SAT and $Y$ is the predictor random variable science GPA given by equation $$\hat Y =...
2 votes
1 answer
2k views

Sum of Squared Error Chi-Square distribution degree of freedom in Multilinear Regression

In this link it says that $Y$ variables has zero covariance (because covariance matrix has only diagonal terms) which implies they are independent. Actually in linear regression $Y$ takes its ...
  • 85
0 votes
0 answers
794 views

What is the difference between the residual, lack of fit and pure error In F test for Regression Analysis?

What is the difference between the residual, lack of fit and pure error and how to calculate each of them in ANOVA F test for nonlinear regression?
  • 1
2 votes
1 answer
249 views

Two way ANOVA, no difference between tests based on type I vs type II sums of squares

I'm learning about two way ANOVA possibly with interaction. I'm following this tutorial http://www.sthda.com/english/wiki/two-way-anova-test-in-r This is their code. ...
0 votes
1 answer
21k views

Calculating SSE in R

I'm following the "Intro to Statistics" course in Data Camp and I'm having some trouble as it seems that the course is contradicting itself: https://s3.amazonaws.com/assets.datacamp.com/production/...
0 votes
0 answers
84 views

Why is regression line represented as $y = b0 + b1 * x$?

I am new to Data Science and ran into Regression Line formula which is $ y = b0 + b1 *x $ (where x is dependent variable, y is predicted variable) I understood the meaning of this formula as a ...
2 votes
1 answer
743 views

Better to Minimize Absolute Error or Sum of Squared Error?

I have an Excel model which predicts the number of customers for a given month. The prediction depends on a churn rate. I have the absolute error (actual vs predicted), along with squared error and ...
4 votes
1 answer
5k views

Why error sum of squares has n-2 df (possibly not duplicate, please read on)? (Regression Question Series - Part 4)

In simple linear regression, the error sum of squares is given by $$ \text{SSE} = \sum_{i=1}^n(y_i - \hat{y_i})^2 \\ \hat{\sigma}^2 = s^2 = \dfrac{\text{SSE}}{n-2} $$ where $n-2$ is the degrees of ...
1 vote
1 answer
473 views

Minimize Logged Sum of Squares?

When numerically maximizing the likelihood function it is standard practice to do this indirectly by minimizing the negative log-likelihood. When numerically minimizing the residual sum of squares (...
3 votes
0 answers
194 views

R: anova(lm): What is the Sum Sq when we have two inputs

What is the formula to calculate the Sum sq column for the inputs? Answers to some other questions state, that it should be $RSS = \sum (\hat Y_i -\bar Y)^2$, yet this is false, it only gives the ...
  • 723
2 votes
2 answers
2k views

Showing that $\sum_{i=1}^n (y_i-\hat{y_i})(\hat{y_i} - \bar{y}) = 0$ for the generalized linear model [closed]

Exercise : Prove that for the generalized linear model, it is : $$\sum_{i=1}^n (y_i-\hat{y_i})(\hat{y_i} - \bar{y}) = 0$$ Question : How would one proceed with proving that for the generalized ...
  • 173
1 vote
1 answer
59 views

Squaring floats between -1 and 1 reduces sum of squares, so why do it? [duplicate]

I have been learning basic statistical testing as it relates to agriculture and have become familiar with the common practice of summing squared raw deviation values, whether in something simple like ...
  • 113

1
2 3 4 5