Logic regression, also known as logic regression analysis, is a generalized linear regression analysis model, commonly used in data mining, automatic diagnosis, economic forecasting, etc.
Logic regression is linear regression that begins with the output result as a continuous value of the actual meaning, and therefore there is a lot of lack of linear regression analysis.
Logical regression model
Logic regression is an ultimate model, which corresponds to y = f (x), indicating the relationship between the variable X and the variable y. The most common questions, such as: Wonderful thinking when he is sick, and then judge whether the patient is sick or giving birth to something, and what is the understanding of the self-variable X, the characteristic data, and the determination is equivalent to obtaining the variable y, predictive Classification.
Figure 1 Linear regression example
The simplest return is linear regression. For the explanation of Andrew NG, as shown in Figure 1.a, X is the data point - the size of the tumor, Y is the observation value --- whether there is a malignant tumor. By constructing a linear regression model, such as Hθ (X), after constructing the linear regression model, it can be either dependent on the tumor size, and whether it is malignant tumor Hθ (x) ≥ 0.5 is malignant, hθ (x) <0.5 is="" benign.="" at="" the="" same="" time,="" the="" robustness="" of="" linear="" regression="" is="" very="" poor,="" such="" as="" establishing="" regression="" on="" the="" data="" set="" of="" figure="" 1.b,="" due="" to="" the="" existence="" of="" the="" rightmost="" noise,="" making="" the="" regression="" model="" in="" the="" training="" set.="" this="" is="" mainly="" due="" to="" the="" consistent="" linear="" regression="" in="" the="" entire="" real="" domain,="" and="" the="" scope="" of="" the="" classification="" needs="" to="" be="" [0,="" 1].="" logic="" regression="" is="" a="" reducing="" prediction="" range,="" defining="" a="" predicted="" value="" of="" [0,1],="" with="" regression="" equation="" and="" regression="" curve="" as="" shown="" in="" fig.="" when="" the="" logic="" curve="" is="" in="" z="0," it="" is="" very="" sensitive,="" in="" z=""> 0 or z <0, is="" not="" sensitive,="" and="" the="" predicted="" value="" is="" defined="" as="" (0,="" 1).="" figure="" 2="" logical="" equation="" and="" logic="" curve="" logic="" returns="" to="" actually="" uses="" a="" logical="" function="" based="" on="" linear="" regression,="" but="" also="" because="" of="" this="" logic="" function,="" logic="" returns="" to="" a="" dazzling="" star="" in="" the="" field="" of="" machine="" learning,="" but="" also="" the="" core="" of="" the="" advertising,="" for="" multi-logical="" logic="" return,="" the="" following="" formulas="" are="" similar="" and="" classified,="" in="" which="" the="" transformation="" of="" the="" formula="" (4)="" will="" be="" estimated="" at="" the="" logic="" regression="" model="" parameter,="" and="" the="" simplification="" formula="" has="" brought="" a="" lot="" of="" benefits,="" y="{0," 1}="" is="" a="" classification="" result.="" 2.="" decide="" the="" boundary="" why="" is="" logical="" regression="" to="" solve="" the="" classification="" problem?="" we="" can="" use="" the="" judgment="" boundary="" to="" explain,="" it="" is="" understood="" that="" the="" boundary="" of="" different="" categories="" of="" data="" is="" used,="" and="" both="" of="" the="" boundaries="" should="" be="" data.="" from="" the="" two-dimensional="" right="" angle="" coordinate="" system,="" several="" examples="" are="" probably="" the="" following="" three="" types:="" from="" the="" above="" three="" graphs,="" the="" red-green="" sample="" is="" a="" sample="" of="" different="" categories,="" while="" we="" are="" drawn,="" whether="" it="" is="" a="" straight="" line,="" circle="" or="" curve,="" it="" is="" better="" to="" separate="" the="" two="" types="" of="" samples="" in="" the="" figure.="" is="" the="" judgment="" boundaries="" we="" said,="" how="" is="" the="" logic="" regression="" to="" get="" these="" decision="" borders="" according="" to="" the="" sample="" point?="" we="" still="" tell="" this="" problem="" in="" some="" examples="" of="" the="" approach="" of="" professor="" andrew="" ng.="" go="" back="" to="" the="" sigmoid="" function,="" we="" found="" that="" when="" g="" (z)="" ≥="" 0.5,="" z≥0;="" for="" hθ="" (x)="g" (θtx)="" ≥="" 0.5,="" θtx="" ≥="" 0,="" this="" point="" meant="" y="1;" contrary="" when="" predicted="" y="0," θtx<0;="" so="" we="" believe="" θtx="0" is="" a="" decision="" boundary,="" and="" when="" it="" is="" more="" than="" 0="" or="" less="" than="" 0,="" the="" logical="" regression="" model="" predicts="" different="" classification="" results,="" respectively.="" first,="" first,="" the="" first="" example="" hθ="" (x)="g" (θ0="" +="" θ1x1="" +="" θ2x2),="" where="" θ0,="" θ1,="" θ2="" are="" taken="" -="" 3,="" 1,="" 1,="" respectively.="" when="" -3="" +="" x1="" +="" x2="" ≥="" 0,="" y="1;" then="" x1="" +="" x2="3" is="" a="" decision="" boundary,="" the="" graphic="" is="" as="" follows,="" just="" open="" the="" two="" types="" of="" points="" on="" the="" map:="" example="" 1="" is="" just="" a="" linear="" decision="" boundary.="" when="" hθ="" (x)="" is="" more="" complicated,="" we="" can="" get="" nonlinear="" decision="" boundaries,="" such="" as:="" at="" this="" time,="" when="" x12="" +="" x22="" ≥="" 1,="" we="" determine="" y="1," at="" this="" time,="" the="" decision="" boundary="" is="" a="" circular,="" as="" shown="" below:="" so="" we="" found="" that="" in="" theory,="" as="" long="" as="" our="" hθ="" (x)="" design="" is="" reasonable="" enough,="" it="" is="" accurate="" to="" say="" that="" θtx="" in="" g="" (θtx)="" is="" sufficiently="" complex,="" we="" can="" fit="" the="" boundary="" of="" different="" judgment="" borders="" in="" different="" situations,="" thus="" different="" sample="" points="" are="" separated.="" intuitively="" understand="" the="" logic="" regression="" in="" the="" two-dimensional="" space,="" which="" is="" the="" characteristics="" of="" the="" singmoid="" function,="" so="" that="" the="" threshold="" of="" the="" determined="" threshold="" can="" be="" mapped="" to="" a="" planar="" decision="" boundary.="" of="" course,="" as="" the="" characteristics="" are="" complicated,="" it="" is="" determined="" that="" the="" boundary="" may="" be="" a="" variety="" of="" samples,="" but="" it="" can="" better="" separate="" the="" two="" types="" of="" sample="" points="" to="" solve="" the="" classification="" problem.="" ,="" read="" the="" full="" article,="" original="" title:="" machine="" learning="" |="" one="" article,="" what="" you="" read,="" what="" is="" logic="" article="" source:="" [micro="" signal:="" dkiot888,="" wechat="" public="" number:="" dingku="" iot="" tribe]="" welcome="" to="" add="" attention!="" please="" indicate="" the="" source="" of="" the="" article.="">
Our other product:
